It saves time and lives.
A recent study suggests that individuals with breathing difficulties could be diagnosed with chronic obstructive pulmonary disease (COPD) through a new AI-assisted lung test. This AI technology can accurately detect COPD from a single chest CT scan taken during inhalation, according to findings published on December 12 in Radiology: Cardiothoracic Imaging.
In addition to diagnosing COPD, the AI is capable of assessing the severity of the condition based on the scan. Previously, diagnosing and staging COPD required two CT scans—one during inhalation and another during exhalation. However, this new approach streamlines the process with a single scan combined with relevant clinical data.
“Our study demonstrates that diagnosing and staging COPD can be accomplished with just one CT scan and the appropriate clinical information,” said Kyle Hasenstab, assistant professor of statistics and data science at San Diego State University.
COPD, a group of progressive lung diseases, significantly hampers breathing and is the third leading cause of death globally. Traditionally, COPD has been diagnosed using spirometry tests that measure lung function by evaluating the amount of air inhaled and exhaled. In some cases, hospitals also use CT scans to identify lung damage contributing to breathing difficulties, though this practice isn’t universally adopted due to the expertise required for imaging and interpretation.
Researchers believed that integrating AI could make CT scans more accessible for COPD diagnosis, reducing the need for specialized training to analyze the images. For this study, the team used CT scans and spirometry data from nearly 8,900 patients, with an average age of 59, all of whom had a history of smoking.
The AI system was trained to predict spirometry measurements from the CT scans and clinical data. The results showed that the AI could accurately diagnose COPD with just one CT scan, whether it was taken during inhalation or exhalation. The addition of clinical data further enhanced the AI’s diagnostic accuracy.
According to Hasenstab, this method could not only increase access to COPD testing but also reduce patient costs, discomfort, and exposure to harmful radiation.
Discussion about this post