5 May 2020

Researchers from King’s College London School of Biomedical Engineering and Imaging Sciences will be commencing a COVID-19 project to help clinical staff more effectively triage patients with COVID-19.

Researchers at King’s College London School of Biomedical Engineering and Imaging Sciences and King’s College Hospital NHS Foundation Trust will be commencing a COVID-19 project using chest X-rays, clinical data and AI to help clinical staff more effectively triage patients with COVID-19 and the likelihood of needing ventilator support.

The chest X-ray is a diagnostic imaging workhorse of the NHS. Almost all patients with suspected COVID-19 will have a chest X-ray.

In this study, up to 10,000 adults with suspected COVID-19 receiving care at King's College Hospital NHS Foundation Trust will receive chest X-rays and be monitored over several months.

Data collected during this time will be analysed and used to inform machine learning approaches in order to create an AI ‘decision-support’ tool.

This tool can then be used by clinical staff to help them determine which patients, based on the data, will benefit the most from a ventilator.

Lead researcher and consultant neuroradiologist at King’s College Hospital NHS Foundation Trust Dr Thomas Booth said that, if proven, this tool will enable a simple chest X-ray to predict the outcome that ventilation will have on each patient:

Clearly, if we can extract additional information from the chest radiograph, that predicts whether a ventilator can help the patient, then that would help determine which patients will benefit the most from mechanical ventilation.

Recent evidence from abroad has focused on changes in computed tomography (CT) scans of the chest, for example documenting CT changes over time or determining an algorithm that can differentiate COVID-19 pneumonia from other pneumonia on CT scans.

However, some hospitalised COVID-19 patients are undergoing less CT scanning in the NHS as departmental resources have been redirected to prioritise chest radiograph reporting over any other imaging investigation.

Dr Booth continues:

Importantly, an algorithm developed now could also be used if there are recurrent or seasonal COVID-19 peaks.
More broadly, AI methods do allow what is learnt from one disease group to be used in another.

The project is receiving funding from BCB Group as part of their BCB Seed Foundation, which is currently focusing on ways to combat the COVID-19 pandemic.

Founder and CEO Oliver von Landsberg-Sadie from the funders BCB Group said this pioneering research exemplifies our belief in the power of innovation to solve the most pressing issues of today, so that together we can create a better future. He said:

To support both those working on the front line and across vulnerable areas of society during this time, we are honoured that BCB Group’s Seed Foundation and our clients are supporting the researchers involved in this high-impact project which has the potential to deliver positive impact not only to the front line, but also to the global community beyond.

Our allergy, respiratory, critical care, anaesthetics and pain Clinical Academic Group has world class specialist clinics in interstitial lung disease, sleep, cystic fibrosis, bronchiectasis, lung cancer, chronic obstructive pulmonary disease and asthma.