Optimization of Case-Specific Vascular Tree Models Based on Vessel Size Imaging

We analyze a problem, which is relevant for physiological modeling of vascular networks: the initialization and optimization of specific, individual models of a functioning vessel network. We do not try to describe the process of growing vessels. Rather, the model properties are optimized via similarity metrics between the artificial vessel network and suitable image data. A vascular network model has many degrees of freedom to optimize and it can be assumed that there are many different optimal solutions. In order to reduce the variability between different solutions we enforce certain physiological properties, for instance Murray’s Law, which must be fulfilled by an optimal network. We propose and validate several similarity metrics, which can be used to optimize the radii of the vascular tree.