A new ultrasound test has been developed in London that can diagnose COVID-19 through patterns in ultrasound images.
Researchers at Lawson Health Research Institute (Lawson) developed and tested which uses artificial intelligence, trained to learn and recognize patterns in ultrasound lung scans of patients with confirmed COVID-19 infection.
The scans were compared to images of other types of lung diseases and infections, both of which “produce a highly abnormal imaging pattern.”
“We sought to determine if there was something beneath the surface that the computers could figure out,” Dr. Robert Arntfield, Lawson researcher and medical director of the Critical Care Trauma Centre at London Health Sciences Centre (LHSC).
Arntfield said different lung diseases in images are not noticeable to the naked eye but their new diagnostic technique can decipher the difference.
“There are details that distinguish COVID-19 at the pixel level,” Arntfield said.
“It’s almost like the AI sees a QR code that we cannot see, unique to the disease.”
The team at Lawson studied more than 100,000 ultrasound images from patients with confirmed COVID-19 infection at LHSC and ultrasound scans of patients with other types of lung diseases and infections.
Although ultrasound has been around for some time, Arntfield said the use of it on the lungs is relatively new, only being used in the last 20 years or so, and has become an important tool for the diagnosis and care of critically ill patients experiencing acute respiratory failure.
Arntfield said because the field of AI is heavily governed it can’t be immediately used to diagnose patients but he’s optimistic of the promising future for the portable low-cost machine which can be used to diagnose not only COVID-19 but other lung infections as well.
“What is in our immediate environment that we can’t see but is visibly available to an algorithm?” he said.
“The ability to apply ultrasound and ingest those images in real time to give predictions about what you have or don’t have in your lungs is a strong chance to enrich health care.”
Lawson has recently approved Arntfield’s Project Deep Breathe which aims to go beyond COVID-19 and explore multiple conditions where lung ultrasound and AI can be paired.
The study, Development of a convolutional neural network to differentiate among the etiology of similar appearing pathological B lines on lung ultrasound: a deep learning study, is published in BMJ Open.