Research On Optimal Angiographic Viewing Angles For The Segment Of Interest And Arterial Motion Estimation Based On MSCT

Coronary artery disease (CAD) is a major threat to poeple's health in Chima. Among the invasive and nonin-vasive diagnostic methods, X-ray coronary angiography(CAG) and multislice computed tomography(MSCT) are the most popular. Though regarded as gold standard for CAD, CAG is essentially two-dimensional(2D) projection of three-dimensional(3D) vessel structures. In unsuitable viewing angle, the vessel projection for segment of interest may have serious foreshortening or overlapped by other vessels, which can influence the assessment and treatment of lesions. Therefore, it is critical to determine the optimal viewing angle.As a rapidly evolving 3D noninvasive imaging method, MSCT has been widely used in CAD diagnosis and pre-operative planning. The main work of this thesis is to analyze the MSCT cardiac data with computer vision and digital image processing knowledge, including: 3D coronary artery segmentation, CAG system geometric model building up, calculation of optimal angle for segment of interest in interventional operation, multi-phase coronary motion analysis. Such results are essential to assist cardiologists in planning and performing a percutaneous coronary intervention (PCI).Major work of this thesis can be included in four parts as follows:1) Design of the the whole schemeProposed the the whole scheme for pre-operative plan based on MSCT data analysis acoording to MSCT and CAG system parameters.The data processing flow is designed to calculate the optimal angiographic viewing angles and analyze coronary motion.2) 3D coronary segmentation in MSCT cardiac dataRegion-growing extraction scheme is utilized and some necessary pre-processing measures are combined, such as probabilistic decision, Hessian-matrix-based vessel filtering and optimal-oriented-flux-based vessel filtering to obtain satisfactory results.3) Calculation of optimal angiographic viewing angles based on MSCT vessel modelGeometric model is built up for coordinate transformation from MSCT data coordinate system to CAG system coordinate system. A perspective projection was performed to obtain the 2D simulated angiogram. After 3D morphologic thinning and B-spline fitting for segment of interest, the optimal viewing angle is calculated with the minimum foreshortening and minimum overlapping principle.4) Coronary motion analysis based on MSCT vessel dataMulti-phase skeleton point non-rigid registration is performed based on coherent point drift (CPD) algorithm to estimate the motion field. Quantitative description for displacement amplitude and velocity of main branches is given. Dynamic coronary model is built up with the full-phase cardiac data.Optimal angiographic viewing angles in the whole cardiac cycle are calculated.