AI can now detect COVID-19 in lung ultrasound images.

Artificial intelligence can spot COVID-19 in lung ultrasound images, just as facial recognition software can spot faces in a crowd, new research suggests.

The findings advance AI-powered clinical diagnostics and bring healthcare professionals closer to being able to quickly diagnose patients with COVID-19 and other pulmonary diseases with algorithms that predict disease progression. Comb through ultrasound images to identify symptoms.

The results, published in New Communication Medicineculminating in an effort that began early in the pandemic when clinicians needed tools to rapidly evaluate legions of patients in overwhelmed emergency rooms.

Senior author Moinato Bell, John C. Malone Associate, said, “We developed this automated detection tool to help physicians in emergency settings quickly and accurately diagnose a large number of patients. are needed, as in the early stages of an epidemic”. Professor of Electrical and Computer Engineering, Biomedical Engineering, and Computer Science at Johns Hopkins University. “Potentially, we want to have wireless devices that patients can use at home to monitor the progress of COVID-19.”

Co-author Tiffany Fong, assistant professor of emergency medicine at Johns Hopkins, said the tool also has the potential to lead to wearable devices that detect diseases such as heart failure, which cause fluid overload in patients’ lungs. can become, not unlike COVID-19. Medicine.

“What we’re doing here with AI tools is the next big frontier for care,” Fong said. “An ideal use case would be a wearable ultrasound patch that monitors fluid build-up and tells patients when they need medication adjustments or when they need to see a doctor.”

AI analyzes ultrasound lung images to spot features known as B-lines, which appear as bright, vertical abnormalities and indicate inflammation in patients with pulmonary complications. It combines computer-generated images with real ultrasounds from patients—including some who received care at Johns Hopkins.

“We had to model the physics of ultrasound and sound wave propagation well enough to get reliable simulations,” Bell said. “We then took it a step further to train our computer models to use these simulated data to reliably interpret real scans of patients with affected lungs.”

Bell said that early in the pandemic, scientists struggled to use artificial intelligence to predict signs of COVID-19 in ultrasound images of the lungs because of a lack of patient data and because they They were just beginning to understand how this disease manifests itself in the body.