Author: Yik Ching Joshua Lee

Publisher:

Published: 2015

Total Pages: 263

ISBN-13:

Acute pulmonary embolism (APE) has a high mortality and many cases of APE go undiagnosed, as the pulmonary circulation is relatively hidden from clinical examination. The pathophysiology of APE is not completely understood, as there is a complex interplay of mechanisms that contribute to the disorder’s response. A difficulty in treating APE is that the mechanisms contributing to response are not well defined, and therefore it is difficult to predict which patients will respond most sensitively to a given clot load based on clinical evidence. Insight into the mechanisms of APE progression and severity has relied on controlled animal studies. Pigs are a widely-used experimental animal for representing human physiology and pathophysiology, because their comparative anatomy, as well as physiological and pathophysiological responses, are said to closely resemble that of humans. However, differences between pig and human in size and lung anatomy leads to translational limitations that are sometimes overlooked. Computational models with appropriate validation could bridge the gap in translating data from animal studies to human clinical practice. In the area of APE this translation is currently limited by a lack of a validated structure-function model for perfusion of the porcine lung. The branching geometry of the pulmonary arterial and venous trees in pig is different in structure to the human pulmonary vasculature, and studies have previously suggested that species-specific branching asymmetry of the pulmonary blood vessels contributes to differences observed in pulmonary blood flow distribution between species. A realistic model that accurately reflects the geometry and mechanical properties of the in vivo porcine lung is therefore critical for translating detailed investigation of structurefunction relationships in the pulmonary circulation of the pig to human. The overall aim of this research was to develop a novel, validated computational model for the porcine pulmonary circulation, that can be used to understand the interplay between the fundamental mechanisms of pulmonary vascular disease. A structure-based theoretical model that integrates new imaging and experimental data, plus previous experimental and clinical observations, is presented here. This thesis presents a quantitative analysis of the pulmonary arteries in five pig lungs, characterising their branching pattern, inter-subject similarity, and self-similarity in branching geometry. A summary model for the self-similar pulmonary arterial tree is described. A method for generating anatomically-based finite element models of the porcine pulmonary vascular tree was developed, based on previous volume-filling branching methods and the new knowledge of the porcine pulmonary arterial tree morphometry. Subject-specific spatially distributed models were generated for each animal using this new method (in the prone posture, at close to full lung expansion), and the full pulmonary arterial tree geometry statistics were compared with experimental data from the five animals. The generated models were consistent with the data with respect to key morphometric parameters of branching angles, rates of reduction of branch diameter and length with branch order, rate of increase of number of branches in an order with reduction in order, ratios of minor or major child diameters to parent diameter, and length to diameter ratios. A multi-scale model was implemented to simulate the distribution of perfusion in the porcine lung. The model includes an approximation for the deformation of the lung tissue due to change in lung size and posture. Model predictions for the lung supine, at close to functional residual capacity, compared well with the haemodynamic data from each animal at baseline. The performance of the model was assessed for predicting haemodynamics and gas exchange following arterial occlusion in APE. The model predicted the general trends of the experimental data, but was not completely consistent with regional functional imaging. The model also suggested that recruitment of small vessels (arterio-venous shunts, or supernumerary vessels) could be important for mitigating increase in pulmonary vascular resistance when the proportion of occluded lung increases. An important question was whether a subject-specific model is necessary for all studies, or whether a single (generic) geometry with appropriate boundary conditions is sufficient to reproduce the important behaviours of the pulmonary circulation. A generic species-specific model was therefore developed and validated, by demonstrating that any subject-specific porcine model can be parameterised to reflect individual pulmonary vascular function that has been measured for any other subject. The model was extended further by including a model for hypoxic pulmonary vasoconstriction. Simulation of normoxic and hypoxic ventilation was compared against experimental data from an independent study. The model prediction of arterial constriction during hypoxia (indicated by elevation of pulmonary artery pressure) and change in blood gases from normoxia were consistent with experiment. This research has established a new validated model to complement animal experimental studies, such that the interaction of mechanisms that contribute to APE can be investigated and presented in a quantitative way.