NHSX develops national Covid-19 image database to train AI

Hospitals and universities across the country can now access thousands of Covid-19 images and scans in a bid to develop artificial intelligence solutions to tackle the virus.

NHSX has collected more than 40,000 CT scans, MRIs and X-rays from more than 10,000 patients across 18 NHS trusts over the course of the pandemic.

Together they form the National Covid-19 Chest Imaging Database (NCCID) and have been extended to hospitals and universities who are using the images to track patterns of illness.

It is hoped the database will speed up diagnosis of coronavirus, ultimately leading to quicker treatment and less pressure on the NHS by predicting things like the need for additional ICU capacity.

The British Society of Thoratic Imaging, Royal Surrey NHS Foundation Trust and AI company Faculty are working with NHSX on the database as part of the NHS AI Lab.

The database could help inform the development of a national AI imaging platform to safely collect and share data, used to develop AI technologies to address a number of other conditions such as heart disease and cancers, according to a government statement.

Dominic Cushnan, head of AI imaging at NHSX, said: “We are applying the power of artificial intelligence to quickly detect disease patterns and develop new treatments for patients. There is huge potential for patient care, whether through quicker analysis of chest images or better identification of abnormalities.

“The industrial scale collaboration of the NHS, research and innovators on this project alone has demonstrated the huge potential and benefits of technology in transforming care.”

Clinicians at Addenbrooke’s Hospital in Cambridge are currently using the NCCID to develop an algorithm to inform a more accurate diagnosis of patients when they present with potential Covid-19.

Using visual signatures of the virus, as they appear in chest scans, they are able to compare the patterns in the patient’s imaging with those seen previously in the NCCID to get a more accurate diagnosis and prognosis.

Carola-Bibiane Schönlieb, head of the Cambridge Image Analysis group at the University of Cambridge, added: “The NCCID has been invaluable in accelerating our research and provided us with a diverse, well-curated, dataset of UK patients to use in our algorithm development.

“The ability to access the data for 18 different trusts centrally has increased our efficiency and ensures we can focus most of our time on designing and implementing the algorithms for use in the clinic for the benefit of patients.

“By understanding in the early stages of disease, whether a patient is likely to deteriorate, we can intervene earlier to change the course of their disease and potentially save lives as a result.”

Researchers from universities in London and Bradford are also using the database to develop AI tools that could help doctors improve the treatment for patients with Covid-19.

The NHS AI Lab recently launched a £140m AI award which aims to bring the benefits of the technology into the health and care sector.