Diagnosing COVID-19 using artificial intelligence

Researchers at Lawson Health Research Institute are investigating whether an artificial neural network could be used to diagnose COVID-19. The AI system is being trained to learn and recognize patterns in ultrasound lung scans of patients with confirmed COVID-19 at London Health Sciences Centre (LHSC) by comparing them to ultrasound scans of patients with other types of lung infections.

“Machines are able to find patterns that humans cannot see or even imagine. Lung ultrasound scans of patients with COVID-19 pneumonia produce a highly abnormal imaging pattern.  This pattern isn’t unique to COVID-19, and can be seen in other causes of pneumonia.  It is plausible, however, that there are details that distinguish COVID-19 at the pixel level that cannot be perceived by the human eye,” explains Dr. Robert Arntfield, Lawson Researcher and Medical Director of the Critical Care Trauma Centre at LHSC.

“If we can train a neural network to learn and identify these unique characteristics among different scans, we can apply this AI to enhance the diagnostic power of portable ultrasound.”

Point-of-care ultrasound has become increasingly important for the care of critically ill patients and LHSC is a global leader in the use of this technology at the bedside. Lung ultrasound has proven to be effective in diagnosing different types of lung infections and illnesses, such as pneumonia, with a high degree of accuracy. The convenience, portability and low cost of using these machines has helped them become a standard bedside tool in emergency departments and intensive care units worldwide.

This research project is part of a grassroots effort by a small group of local clinicians to innovate and create technology to solve sophisticated problems with AI. With many of Dr. Arntfield’s team having a background in computer programming, they were able to code the neural network being tested. Minimal funding was required, with the project being driven largely by the urgency of COVID-19 coinciding with the recent creation of this clinical AI working group.

“Our research team has used AI to help improve diagnostics related to other parts of the body. This project is a great example of the unique ability we have here in London to be agile: that is, to identify a gap and move quickly towards finding a solution,” says Dr. Arntfield. “I am thrilled that we were able to move through the approval process quickly, and get our ideas working in such a short amount of time.”