Oxford to collaborate on AI algorithms to help predict COVID-19 severity in patients

The University of Oxford-led National Consortium of Intelligent Medical Imaging (NCIMI) in the UK is collaborating with GE Healthcare to develop and test algorithms to aid in the diagnosis and management of COVID-19 pneumonia.

The program will focus on developing, enhancing and testing potential algorithms to help diagnose COVID-19 pneumonia, predict which patients will develop severe respiratory distress – a key cause of mortality in patients who develop COVID-19 pneumonia – and which patients might develop longer term lung function problems, even when they recover from respiratory distress.

At present, clinicians cannot easily predict which patients who test positive for COVID-19 will deteriorate and require hospital admission for oxygen and possible ventilation. Nor is it clear which patients will suffer long-term consequences from the lung damage from COVID-19 pneumonia. The teams aim to develop algorithms incorporating data from thousands of patients medical imaging, laboratory and clinical observations to provide both a quicker diagnosis and a prediction of how a patient may progress and recover.

Currently, some patients admitted to hospital do not see a worsening of their symptoms, while others who appear stable can deteriorate rapidly. Identification of those patients at highest risk of deterioration and long-term lung function problems may help physicians and caregivers to accelerate intensive support. It may also allow those with lower risk to be monitored in a suitably safe environment, potentially including the patient’s home. GE Healthcare and NCIMI aim to develop tools to help in the management of these COVID-19 patients from triage to acute monitoring, interventions, to discharge and those requiring follow-up after recovery.

“It would be extremely valuable to predict at a relatively early stage in the disease which patients will do well, which are at risk of imminent deterioration and should be admitted to ICU as they will need more intensive support, and which are at higher risk of delayed deterioration and need to be actively monitored.” says Professor Fergus Gleeson, Consultant Radiologist, Professor of Radiology at the University of Oxford, and the 2020 president of the European Society of Thoracic Imaging. “These distinctions would allow hospital resources to be targeted to those that will need them whilst in hospital and following discharge.”

“As health systems manage COVID-19 cases, clinicians can benefit from new technologies to help triage and determine which patients are likely to develop respiratory distress and longer-term complications,” said Kieran Murphy, President and CEO of GE Healthcare. “If we can ensure patients are quickly placed in the right care setting, this may help to improve outcomes.”

The development of robust algorithms and models requires large data sets comprising thousands of patients. The Oxford and NCIMI teams will have access to data from NCIMI NHS partner hospitals as well as working with the National COVID-19 Chest Imaging Database (NCCID) led by NHSX in England and the British Society of Thoracic Imaging. GE Healthcare is developing various imaging and vitals-sign algorithms for use in conducting research for better understanding of the COVID-19 disease progression. The team at Oxford will assess and test various approaches to determine if these can be used to help patients who have or have had COVID-19 pneumonia.

/University Release. View in full here.