Simulating Blood Flow in an Anatomical Arterial Network

Abstract

The overall aim of this thesis is to develop a computational model of the human arterial system to predict blood pressure and ow rates throughout the body. The model used an anatomical vascular network derived from the Visible Human Project and implemented in OpenCMISS, an open-source mathematical modelling environment that enables the application of nite-element analysis and other techniques to a variety of complex bioengineering problems. This arterial model coupled a one-dimensional arterial tree model with a lumped parameter description of the peripheral circulation is capable of simulating global hemodynamics and arterial wave propagation simultaneously and can provide the downstream boundary conditions for three-dimensional simulations at the particular arterial location using one-dimensional models of the entire arterial system and zero-dimensional models at the distal ends. The model provided better predictions of published experimental measurements compared to existing one-dimensional models. The model was used to simulate pressure and ow waveforms in pathological conditions to investigate potential uses in surgical planning and the use of inexpensive ow and pressure measurements for diagnosis. This includes the use of the model to explore the possibility of non-invasive disease diagnosis and comparison of di erent surgical options. The ow model was then coupled with an advection-di usion equation to simulate the transport of solutes (drugs, hormones) through the vasculature. For the rst time, by coupling with lumped parameter descriptions of metabolism and excretion in target organs, the extended model provides a physiologically based pharmacokinetic model that will be useful in drug development and regulatory toxicology to predict the kinetics and metabolism of substances in the body. The coupled framework has the ability to model pharmacokinetic response to the administration of drugs in a more physiologically realistic manner compared to the traditionally used compartmental models.An important contribution to the state of knowledge is the free availability of the codebase. Source code and anatomical data are not available for any of the existing models which makes it impossible to independently validate published results and to modify or extend the model for di erent applications. By providing this capability the thesis provides a useful computational resource to the bioengineering community.