Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Medical Biophysics


Christopher G. Ellis


The purpose of this thesis was to examine how the use of real 3-dimensional (3D) capillary network geometries affect models of oxygen transport to tissue. Software was developed to reconstruct microvascular geometry in 3D from intravital video. Characterization of 3D reconstructions demonstrated that capillary density, length and capillary diameter were consistent with previous findings. Using reconstructed capillary networks a strategy was devised that utilized red blood cell (RBC) supply rate (SR) as a metric for flow modeling. Applying the RBC SR based flow model on baseline and perturbed flow conditions demonstrated that RBC SR is a major determinant of oxygen delivery that is insensitive to changes in flow distribution. The resulting flow solutions were used for comparing oxygen transport in 3D networks and synthetic parallel arrays. A variety of physiological conditions were simulated and it was determined that parallel arrays resulted in oxygen transport solutions with higher mean PO2 due to homogeneous distribution of vessels in the volume. Lastly, to investigate oxygen transport in a complex pathology a model of sepsis was used to investigate how incremental perfusion loss, consumption increase and change in RBC SR affect oxygen delivery. It was shown that perfusion loss did not markedly impair oxygen delivery provided that RBC SR was maintained. These results have improved our understanding of oxygen transport to tissue in normal and diseased conditions; the use of reconstructed networks and measurements of blood flow & oxygen saturation in computer models provides different solutions than those using statistical averages and synthetic networks.