Ph.D Thesis

Ph.D StudentKaimovitz Benjamin
SubjectA Full 3-D Reconstruction of the Entire Porcine
Coronary Vasculature
DepartmentDepartment of Biomedical Engineering
Supervisor PROFESSOR EMERITUS Yoram Lanir
Full Thesis textFull thesis text - English Version


The function of the coronary tree is to convey blood to the myocardium, to distribute blood over the capillary bed, and to regulate coronary flow. It is widely acknowledged that the distribution of coronary blood is heterogeneous. Flow heterogeneity is enhanced under pathologies such as ischemia which induces loss of tone in the arterioles thus contributing to flow in-homogeneity. The architecture of the coronary network, its stochastic nature and branching patterns, are corner stone determinants of the coronary perfusion and oxygen supply and their distributions, both local (fractal dimensions) and global (transmural). Accordingly, realistic analysis of the flow distribution must be carried out based on a three-dimensional (3-D) representation of the network geometry and its biological variability. Previous reconstruction attempts fell short in accounting for the realistic full scale (order of 109 vessels) network structure. In the present study, a realistic large-scale, stochastic 3-D geometrical reconstruction of the porcine coronary network was developed based on detailed morphometric statistical data and on observed specific network characteristics. The model covers the arterial, capillary and venous portions of the coronary vasculature to form a unified, large-scale morphometric model of the entire coronary network which spans the full spectrum of the vessels’ diameters down to the capillary level. The reconstruction was formulated as large-scale optimization process subject to both global constraints relating to the location of the larger vessels and to local constraints of measured branching statistics. The 3-D structures of the epicardial sub-networks for both the arterial and venous trees, and for the arterial transmural and perfusion sub-networks, were generated by means of global large-scale geometrical optimization by applying Simulated Annealing probabilistic search technique. The unified reconstructed coronary arterio-venous network conforms to the gross anatomy of coronary sub-networks in terms of structure and location of the major vessels. In addition, it manifests close proximity to the measured morphometric statistics of native coronary networks. Hemodynamic analysis in the coronary vasculature which was based on the current 3-D reconstruction predicted relative flow dispersion with fractal dimension which demonstrated good agreement with experiments results.  The unified artero-venous coronary model establishes a requisite foundation for a detailed, realistic large-scale coronary flow analysis that can serve for assessment of longitudinal and spatial distribution of flow and pressure, to uncover the role of the network geometry, and to provide details of local flow fields for evaluation of patho-physiological effects as well as influence of drugs and other treatments.