Researchers use artificial intelligence to detect COVID-19 in hospital patients

Researchers are using artificial intelligence to detect COVID-19 from chest x-rays of hospital patients faster than traditional tests. (University of Minnesota and M Health Fairview / FOX 9)

Researchers at the University of Minnesota and M Health Fairview are developing a new technology that uses artificial intelligence to help doctors detect COVID-19 in hospital patients.

“It does feel maybe science fiction-y, but this is probably going to be a part of our new normal,” said Dr. Genevieve Melton-Meaux of M Health Fairview.

Dr. Melton-Meaux explains artificial intelligence (AI) is having some major impacts on health care in 2020.

“There is a lot of enthusiasm around it right now and if you look over the last 2 - 3 years, the uses for AI in medicine is going up exponentially,” said Dr. Melton-Meaux.

She and her team created an AI program that uses chest X-rays to identify if a patient likely has COVID-19. When patients arrive in the emergency department with suspected COVID-19 symptoms, it’s standard protocol for clinicians to order a chest X-ray. The algorithm automatically evaluates the X-ray and recognizes patterns in the lungs associated with COVID-19. This method can help medical teams learn if a patient has COVID-19 much quicker than traditional testing. 

“If we’re waiting say 5 - 6 hours for a test to come back, we’re able to know sooner that there’s a high likelihood a patient might be positive,” she said.

Researchers used more than 100,000 X-rays to train the artificial intelligence program what to look for in COVID-19 patients. 

Dr. Melton-Meaux says many clinics do not have the ability to do COVID-19 testing right now, but most have access to X-ray equipment. She says this technology can give medical professionals better access to tools to help them recognize if patients have the virus.  

“There are very few centers worldwide that have really developed and implemented these AI-enabled decision support systems,” said Dr. Chris Tignanelli, assistant professor of surgery at the University of Minnesota Medical School.

Dr. Tignanelli, who also helped develop this technology, explains what they taught the program to look for.

“Someone who has COVID, we can see contrast in the lungs, we can see these scattered areas of kind of white contrast in the lungs,” said Dr. Tignanelli.

He says the technology has developed to a point where it’s able to explain to doctors why it came to different conclusions.

“We’re the one teaching the algorithm and now we’re potentially consuming knowledge from that algorithm,” said Tignanelli.

Researchers say that kind of knowledge is helping them make better decisions and better serve patients now and in the future. 

“We’re just at the beginning of it, so I would keep an eye out for it because you’re going to see it more and more,” said Dr. Melton-Meaux.

“I think this is just the tip of the iceberg moving forward,” said Dr. Tignanelli

By working with the medical software company, Epic, this technology has become available for more than 450 health care systems worldwide. 

Researchers at the university with M Health Fairview are also experimenting with a program that takes all of a patient's latest information, such as their vitals, labs and X-rays, and uses data to make predictions about what’s likely to happen with that patient in the coming days. The predictions help the medical team prepare to give the patient what they need or prepare for surges of patients in the ICU.