Artificial intelligence (AI) systems have been used increasingly to diagnose things like breast cancer, skin conditions, and brain tumors. Now, scientists are pioneering a timely use for machine learning. Researchers from New York’s Mount Sinai Hospital believe that they are the first in the United States to use AI to diagnose COVID-19.
The team published a paper in the journal Nature Medicine on Tuesday documenting their findings and explaining how the system works. While AI can’t replace human medical expertise, it could help doctors diagnose COVID-19 earlier and more accurately.
Using Today’s Best Technology
Artificial intelligence is clearly one of the most important technologies of the future. With applications across almost every industry, it is already making an impact.
It has even been helpful in the fight against COVID-19. However, using AI to diagnose the respiratory condition associated with the coronavirus takes things a step further. One of the lead authors of the study, Zahi Fayad, says, “We were able to show that the AI model was as accurate as an experienced radiologist in diagnosing the disease, and even better in some cases where there was no clear sign of lung disease on CT.”
Fayad is the director of the BioMedical Engineering and Imaging Institute (BMEII) at the Icahn School of Medicine. He and other researchers note that lung scans don’t always show signs of disease when a COVID-19 positive patient arrives with initial symptoms. Lab tests to diagnose the disease can take days to come back. The AI system helps physicians address both of those problems.
Despite the fact that COVID-19 is a new problem, the researchers trained their AI model extensively. They note that it practiced on more than 900 scans from medical centers in China, including 419 positive cases and 486 negative ones.
In the United States, CT scans aren’t usually a tool for diagnosing COVID-19. However, researchers believe that it’s time to change that practice. An AI system capable of detecting the disease could help facilitate such a change.
Fayad notes, “The high sensitivity of our AI model can provide a ‘second opinion’ to physicians in cases where CT is either negative (in the early course of infection) or shows nonspecific findings, which can be common.”
He adds, “It’s something that should be considered on a wider scale, especially in the United States, where currently we have more spare capacity for CT scanning than in labs for genetic tests.”
Indeed, testing for the coronavirus has been a hot issue in countries around the world. Despite efforts to ramp up its testing capacities, the U.S. still doesn’t have an ample supply. The AI system could help cut down the need for other forms of tests, especially for patients who are already in the hospital and can get a CT scan readily.
The researchers plan to continue improving the algorithm to make it more accurate. Matthew Levin, director of Mount Sinai Health System’s clinical data science team says, “This is an early proof concept that we can apply to our own patient data to further develop algorithms that are more specific to our region and diverse populations.”