Using AI to assess long-term lung damage due to COVID-19

Area of work


Imaging modality


Size of data set


Dr Fergus Gleeson, NCIMI CMO.

Project lead

Professor Fergus Gleeson

Industry partners

Number of NHS partners


Unmet need

The purpose of this study is to understand in much more detail the effects of COVID-19 on the lungs. We are particularly interested in how long symptoms last and whether we can identify areas on scans that may be the cause of long lasting symptoms.

In some people, the symptoms of COVID-19 can continue for many months after the infection, which may adversely affect their quality of life. For example, many people complain of persistent breathlessness and fatigue. The virus may affect the function of the lung in ways that are not yet fully understood.

Project aims

POST (Post COVID-19 disease follow up imaging using hyperpOlariSed xenon MRI and CT) is reviewing in detail patients from Oxford that have required hospitalisation with COVID-19, and have had a CT scan as part of their clinical care and to follow-up this sub-group of patients with repeat CT scans and hyperpolarised Xenon MRI, HP 129Xe-MRI.

Prior reports on CT follow-up of patients with ARDS have identified a number of CT patterns, most commonly associated with mechanical ventilation. There are specific features of COVID-19 pneumonia that are significantly different to a normal pattern of viral-induced ARDS.

These differences may potentially manifest as persistent radiological changes and clinical symptoms after discharge, and their severity and best mode of management is unknown at present.

Hyperpolarised (HP) 129Xe-MR imaging provides regional information about lung function. HP 129Xe-MRI is able to measure changes in airflow, lung tissue structure and blood flow within specific regions of the lung. The unique ability of 129Xe to dissolve through the alveolar epithelium into the lung parenchyma and capillary blood may enable detection of much smaller changes in post-COVID-19 pneumonia disease perturbation than has previously been possible when investigating post-ARDS changes in pulmonary function.

Co-registering the HP 129Xe-MRI images with the CT images will provide a unique insight into functional and structural changes due to severe COVID-19 pneumonia. If there is a good correlation between disease severity, persistent or slow to improve respiratory function and HP 129Xe-MRI, it may be possible to use this technique in the future avoiding radiation exposure.

Additionally, because HP 129Xe-MRI images gas transfer for the whole lung, and regionally, it may provide insights into why patients are more dyspnoeic than would be suggested by their pulmonary function tests, or potentially identify areas of impaired gas transfer that are normal or near-normal on CT- and therefore explain the relevance of any functional or structural change detected by CT.