Imaging data could identify who should shield in COVID-19 outbreaks

UCLH and UCL researchers are working with Microsoft to see if chest imaging data can identify better who should shield in potential future COVID-19 outbreaks.

With concerns that current shielding guidelines miss out on people who should be shielding but advise others to shield who do not need to, the researchers will be looking at lung and heart scans taken before the epidemic of people who went on to develop COVID-19 and others who did not. The researchers, led by Dr Joseph Jacob and Prof Danny Alexander, will combine this information with demographic data and information on patients’ pre-existing medical conditions, to identify imaging features that can predict severe COVID-19 illness and cardiac events.

The aim is to identify which groups of people are not covered by current shielding guidelines, who should be – and to identify groups where shielding is not necessary.

Data to be analysed is already being collected as part of the National COVID-19 Chest Imaging Database (NCCID), led by Dr Jacob. The UCLH and UCL team will use Microsoft Research’s state of the art Project InnerEye  technology to analyse radiological images.

Dr Jacob, who is Consultant Radiologist at UCLH, a Wellcome Trust Fellow at UCL and academic principal researcher for this project, said: “We are looking forward to carrying out this work with Microsoft Research Cambridge to shed light on who should be shielding – and who does not need to. This work will combine our specialist clinical and imaging knowledge on lung disease and COVID-19 at UCLH and UCL with Microsoft’s machine learning platforms and infrastructure, and image analysis capabilities.”

Read more about the project.