OAK BROOK, Ill. - A deep learning model has demonstrated the ability to accurately diagnose and stage chronic obstructive pulmonary disease (COPD) using just one inhalation lung CT scan, according to a study published in Radiology: Cardiothoracic Imaging.
COPD, a progressive lung disease that severely affects breathing, is the third leading cause of death globally. Traditionally, diagnosis involves a spirometry test that measures lung function through air exchange.
This new approach utilizes a convolutional neural network (CNN) combined with clinical data, allowing for effective COPD diagnosis and staging from a single CT image. The study analyzed data from 8,893 patients, revealing that the CNN model could accurately predict the severity of COPD, classified by the Global Initiative for Obstruct Lung Disease (GOLD) stages.
Dr. Kyle A. Hasenstab, a study author, noted that using a single CT acquisition could enhance accessibility to COPD diagnostics, reducing costs and discomfort for patients while minimizing exposure to radiation.