[Courtesy of Wonkwang University ]
SEOUL -- A computed tomography (CT) scan can be used as an important tool to diagnose COVID-19. However, radiologic diagnostic support is not maintained 24 hours in many institutions and CT may show similar imaging features between COVID-19 and other types of pneumonia, hampering correct diagnosis by radiologists.
A research team led by Lee Jin-seok, a professor from Wonkwang University College of Medicine, has used an AI technique to diagnose COVID-19 pneumonia in CT images and differentiate it from non-COVID-19 pneumonia. Lee conducted a joint study with medical staff from Chonnam National University Hospital and Asan Medical Center in Seoul.
"In patients with COVID-19, symptoms are often mild compared to actual pulmonary lesions, and it's hard to diagnose with simple X-ray images," Lee said in a statement. He said the research team has applied an AI model optimized for diagnosing patients with COVID-19 pneumonia to increase accuracy from about 80 percent to 99 percent.
"This study will help us diagnose COVID-19 patients quickly and accurately, increasing the treatment rate and greatly reducing the burden on medical staff," Lee said. The study was published on the website of the Journal of Medical Internet and Research, a peer-reviewed medical journal.
A simple two-dimensional (2-D) deep learning framework called "Fast-track COVID-19 Classification Network (FCONet)" was developed to diagnose COVID-19 pneumonia based on a single chest CT image, Lee's team said in its research paper, adding that FCONet provides excellent diagnostic performance in detecting COVID-19 pneumonia.
"We described FCONet, a simple 2D deep learning framework based on a single chest CT image, as a diagnostic aid that provides excellent diagnostic performance to diagnose COVID-19 pneumonia," the paper said.
reporter
Lim Chang-won
cwlim34@ajunews.com
