Oxford Biomedical Research Centre has developed an artificial intelligence technology to scan for heightened blood vessel inflammation in COVID-19 patients. The tool can also calculate a COVID-19 patient’s risk of death.
The research, presented at the British Cardiovascular Society Conference, is funded by the British Heart Foundation (BHF) and is one of the six research programmes shortlisted under UK Flagship Projects–a joint initiative of BHF and National Institute for Health Research. UK Flagship Projects are aimed at improving care for COVID-19 patients with heart and circulatory diseases. It provides a framework for the rapid set-up and delivery of high impact COVID-19 research projects across the country.
What’s the need?
Last year, acute cases of COVID-19 had been linked with cytokine storms triggered by the coronavirus. Meaning, immune systems produce proinflammatory cytokines or immunoregulatory cytokines as a reaction to the virus. Patients with high levels of blood vessel inflammation run high mortality risk. High inflammation levels lead to acute respiratory distress syndrome (ARDS) aggravation and tissue damage, ultimately resulting in multi-organ failure, and eventually, death.
However, such patients respond to anti-inflammatory drug Dexamethasone. Thus, if a patient’s cytokines are managed well, it improves their chances of survival.
How does the tech work?
Researchers at the University of Oxford have developed an artificial intelligence platform to identify COVID-19 patients with a high risk of cardiac attacks. They have come up with a COVID-19 signature using machine learning. This particular signature detects any biological red flag in the fat surrounding blood vessels in the chest and measures the cytokine-driven vascular inflammation level in patients.
Charalambos Antoniades, Professor of Cardiovascular Medicine and BHF Senior Clinical Research Fellow at the Radcliffe Department of Medicine at the University of Oxford, wrote the artificial intelligence-powered platform could track vascular disease by decoding information from blood vessel images. The tool will then compare the data with a large RNA bioresource from human tissue biopsies.
Using the artificial intelligence-powered tool, doctors can obtain an inflammatory score. Based on this score, a patient’s risk of death can be measured. These patients, when treated with anti-inflammatory drugs, will have a lesser risk of death. The drug will further help in the long-term recovery process.
In some cases, patients with exaggerated immune responses to the virus have shown high chances of developing abnormal blood clotting. Charalambos said the platform could help in the identification of those patients as well.
The clinical trials are on to test the effectiveness of this approach.
Proof of concept
Researchers at the University of Oxford have applied the COVID-19 signature to CT chest scans of 435 patients at Oxford, Leicester and Bath hospitals. When comparing the degree of inflammation and risk of death in these patients, researchers found the level of cytokine-driven inflammation in the blood vessels was much higher in patients with COVID-19. It was even higher in patients infected by the alpha variant of COVID-19 identified in the UK.
Researchers also established that patients with high levels of blood vessel inflammation were up to eight times more vulnerable and were likely to die in the hospital. On the other hand, patients with vascular inflammation, when treated with Dexamethasone, were found to have a six-fold reduction in risk of dying from COVID-19 compared to those not given the medication.
Researchers at the University of Oxford now plan to study the effectiveness against and impact of COVID-19 variants as and when they emerge. They believe the technology can easily track the long-term cardiovascular effects of COVID-19 and thereby help with quick response to viruses of the future.