Connecting Lung Structure and Function in Cystic Fibrosis Through Physiological Modelling, Image Analysis, and Uncertainty Quantification

The human lung is a very complex organ comprised of millions of airways that range from several centimetres down to micrometers in length. This complexity makes it difficult to reliably relate the outcomes of lung function tests to underlying physical changes in the lung structure. Cystic Fibrosis is a genetic condition that currently affects around 9000 people in the UK. Changes in the lung structure (e.g. airway blockages) of cystic fibrosis patients can indicate progression of an infection associated with the disease that requires treatment. Therefore, catching such changes early is paramount to improving patient care. Currently, spirometry tests (which effectively measure the size and strength of the lungs by monitoring the flow rate during forced breathing through the device) are most commonly used to detect these changes, but these have been shown to be relatively insensitive to the changes associated with early Cystic Fibrosis progression.

The Multiple-Breath Washout test (which involves a patient breathing in a gas mixture through a mask that traces the flow rate and gas concentrations) contains much more information than spirometry. Various formulae and indices exist to summarise lung function from this test data, but it remains unclear how these relate to the underlying structural changes. The goal of this project is to use mathematical and computational modelling to make this connection clearer and develop tools to quantify these changes. I will develop software that can simulate gas flow in lungs with structural changes that are associated with early Cystic Fibrosis. These simulations will be compared directly with data from Multiple-Breath Washout tests and MRI data which images the gas distribution within the lungs as a patient is breathing.

Using the simulations developed we will gain an understanding of the effects of early Cystic Fibrosis progression on these lung function tests, which can be used to develop more sensitive and reliable measures of structural change. Further on, I will develop software tools that can be used to automatically whether a patient's condition is worsening directly from these datasets. Additionally, this project will feed into future research collaborations by expanding these approaches to other disease groups, notably lung transplant patients, COPD and asthma. This project incorporates training in specific expertise in state-of-the-art lung modelling and the necessary advanced mathematical techniques.

The goal of the proposed project is to use advanced computational modelling of gas transport and ventilation in the human airway tree to improve the early detection of Cystic Fibrosis (CF) in both adults and children. This will build on preliminary research employing a novel technique to simulate airway blockages/constrictions (due to mucus 'plugs') accurately, while accounting for uncertainty due to heterogeneity throughout the lung structure without simulating these in detail (greatly reducing the computational requirements). I will use these methods to infer the extent and distribution of lung blockages and the associated uncertainty in these predictions from patient data. I will analyse patient data on multiple breath inert gas washout (MBW), from the NIHR-funded LCI-SEARCH study. Using these studies alongside imaging data (where available) we will synthesise more information into CF diagnosis and early detection through the development and deployment of new clinical indices. I will also develop and incorporate mathematical models of the lung parenchyma that account for physiologically important effects of pulmonary surfactant dynamics and material heterogeneity. Improved measures of early lung physiology will lead to improved detection of disease and more targeted intervention. The lung model developed in CF will also be deployed to investigate early disease interventions in COPD and asthma in future.