Doctoral Student at the
Rex Richards Building, South Parks Road, Oxford OX1 3QU
In the European Union pulmonary disease is the third most common cause of death and in North America the fourth. Currently the two most common forms of assessment of pulmonary function are spirometry and plethysmography. In spirometry the most common parameter of interest is the maximum volume of gas a patient can expire in one second (FEV1), while in plethysmography an estimate of the volume of gas in the thorax is made. To detect the presence of lung disease these measurements are compared to `normal' values for patients of similar physical characteristics. However neither of these techniques is sensitive enough to detect early stage lung disease and both require a significant degree of patient co-operation to perform. Thus there is a need to develop a sensitive method for detecting the early stages of lung disease.
One concept believed to be an extremely important indicator of respiratory health is the degree of inhomogeneity of ventilation. This is usually considered by estimating the ventilation volume distribution of the lungs. This slightly abstract concept can be thought of as sampling small volume elements of the alveoli and determining how much fresh gas enters this volume during each breath. In perfectly homogeneous lungs each unit of volume would receive the same amount of fresh gas, however in practice this is not the case and even healthy lungs display some amount of inhomogeneity. The most common technique used to estimate these distributions is a multi-breath nitrogen washout (MBNW). The procedure involves switching a subject from breathing room air to pure oxygen and monitoring the concentration of nitrogen the patient expires as it washes out their lungs. A ventilation volume distribution compatible with this is then found by varying the fractional ventilations received by 50 compartments of fixed ventilation volume ration so that the data simulated with the model matches that taken experimentally.
However this technique has failed to gain widespread clinical acceptance for two reasons: (i) The poor quality of experimental data and (ii) the mathematical models used to represent this technique are inadequate. We are fortunate to be collaborating with Professor Peter Robbins (Department of Physiology, Anatomy and Genetics, University of Oxford) who has developed experimental apparatus to measure MBNWs with much higher accuracy, thus here we present a novel model (based around the concept that the lung can be modelled using an underlying ventilation volume distribution) that we hope will take advantage of this new more accurate data to provide a clinically useful measure of inhomogeneity of ventilation. My research, based in the Department of Computer Science, is co-supervised by Dr Jonathan Whiteley and Professor Peter Robbins.