Study Says AI Can Enhance Chest CT Assessment of Bronchiectasis in Ever-Smokers

Employing an artificial intelligence (AI) tool to quantify airway-to-artery (AAR) diameter ratios on chest computed tomography (CT), researchers found the percentage of airways with an AAR greater than 1 was associated with increased pulmonary exacerbations in ever-smokers.

Increased pulmonary exacerbations over time have been linked to an increased percentage of airways with airway-to-artery (AAR) diameter ratios greater than 1, according to a new study looking at artificial intelligence (AI) assessment of chest computed tomography (CT) scans in over 4,000 ever-smokers.

Noting that single-section CT assessment of AAR can be time-consuming and limited in the evaluation of airway sections, the study authors compared AI-enabled chest CT and single-section chest CT measurements of AAR as well as visual CT scores of bronchiectasis in 4,192 ever-smokers, including 1,834 people with chronic obstructive pulmonary disease (COPD).

Based on 10-year follow-up data, the researchers found that AI-enabled chest CT measurement of AAR demonstrated an association between a higher percentage of AAR ratios greater than 1 and increased pulmonary exacerbations over time. The study authors also noted that the overall adjusted risk ratio for these patients of 1.08 increased to 1.37 for study participants who had clinical manifestations of bronchiectasis (ranging from cough and dyspnea to phlegm and exacerbation history) and met the imaging criteria for bronchiectasis (> 3 percent of AARs >1). There were similar adjusted risk ratios for study participants who had COPD (1.10) and those with COPD who had clinical and imaging criteria for bronchiectasis (1.32).

The study authors also noted elevated adjusted risk ratios among study participants who had emphysema and met the aforementioned criteria for bronchiectasis (1.38) and those who had chronic bronchitis and bronchiectasis (1.46).

Noting a lack of alignment between visual CT assessment and the AI-based tool, Alejandro A. Diaz, MD, MPH, the lead author of the study, said the AI-based method offers a “continuous metric of the extent of airway dilation on CT scans” whereas current visual and quantitative methods of assessing bronchiectasis do not capture the “complex morphologic changes throughout the bronchial and pulmonary artery trees” seen in ever-smokers.

“The percentage of AAR greater than 1 seemed to capture airway dilation relative to artery size in participants who were not identified as having bronchiectasis at visual inspection,” wrote Dr. Diaz, who is affiliated with the Division of Pulmonary and Critical Care Medicine at the Brigham and Women’s Hospital in Boston, and colleagues. “An interpretation of these findings is that subtle detection of a larger airway size relative to artery size is a difficult and likely variable visual task, compounded by the presence of smoking-related pathologic features at CT, such as emphysema.”

Other potential variability with human assessment could stem from anatomic variability that may contribute to mismatching of the airways and nearby arteries in some lung zones as well as technical failure to recognize mismatches of corresponding generations of airways and arteries, according to the study authors.

“A potential use of an AI-based tool would be as an aid in quickly identifying CT scans with features of bronchiectasis for an in-depth human inspection,” added Diaz and colleagues.

In regard to study limitations, the authors emphasized caution with general extrapolation of the findings given the study’s inclusion of heavy ever-smokers. Diaz and colleagues also acknowledged a lack of validation for the imaging algorithm at histologic analysis and the use of a single AAR cutoff point.