A multi-scale computational model of pulmonary gas exchange

Abstract

The primary role of the lungs - respiratory gas exchange - occurs where air and blood transport meet. Despite the substantial regional variability in both structure and function, the human lungs remain capable of efficient gas exchange by providing close matching of ventilation and perfusion to a region of parenchymal tissue. The goal of this thesis is to develop an anatomically-based computational model that integrates the physical, chemical and mechanical processes that are associated with gas exchange at different spatial scales: pulmonary perfusion, ventilation, gas mixing, and the biochemical reactions in the blood. The models are used to improve our understanding of structure-function relationships by combining anatomical fidelity with accurate biophysical descriptions of these physiological processes, and add to the collaborative Lung Physiome project by providing a gas exchanging model. Two scale-independent models of the blood-gas biochemistry are developed: a detailed model that takes into account the two-phase nature of blood, and a simple reduced model that treats blood as a single-phase fluid. The simple model only incorporates the predominant O2 and CO2 reactions, namely oxyhaemoglobin production and HC03 dehydration. It is shown that under a variety of simulation conditions the simple model is sufficient for simulating 02 uptake but insufficient for CO2 release. The simple gas exchange model is incorporated into a novel model of O2 transport within an acinus, which includes a realistic asymmetric acinar geometry, advective and diffusive O2 transport in the airways, gas exchange flux into the capillary blood, and realistic blood PO2 dynamics. Contrary to previous studies, results from the acinus model show that equilibration between blood and air-side Po2 prevents development of large intra-acinar gradients in alveolar PO2. The theoretical concept of diffusional screening is shown to bear little effect on decreasing acinar efficiency. Furthermore, these results suggest that it is acceptable to assume the acinus is a well-mixed compartment. At the scale of the whole lung, a novel predictive model is developed that integrates spatially heterogeneous tissue deformation, ventilation, perfusion, O2 transport and exchange. The predicted difference between alveolar and arterial PO2 is slightly higher than expected when compared to values in healthy young adults, which suggests that either ventilation heterogeneity is underestimated in the model or active matching mechanisms play a minor but important role in vivo in the homeostasis of ventilation and perfusion distributions. The whole lung gas exchange model is applied to a study of acute lung injury. It is shown that the isolated effect on whole lung gas exchange of a thicker air-blood barrier (which is due to interstitial oedema) is far less detrimental than the effect of stiffer lung tissue, and hence ventilation-perfusion mismatch. Furthermore, the impact of heterogeneous dysfunction is shown to be more detrimental than uniform dysfunction, highlighting the importance of considering the spatial distribution of tissue deterioration. As the lungs are a large organ with complex structures underlying multifarious physiological processes, the anatomically-based integrative model provides a unique tool for further investigations of normal and abnormal lung function.