Background: Exacerbation episodes in individuals with bronchiectasis and COPD are a significant cause of morbidity and healthcare utilization. However, there are scarce data on future exacerbations in coexistent COPD and bronchiectasis. Using an artificial intelligence-based (ai) tool to measure the airway lumen to artery diameter ratios (AAR) at computed tomography (CT) we aimed to assess the association of AARs and future exacerbations.
Methods: We performed ai-based CT measures of AARs in 4,192 smokers with and without COPD. After reconstructing, classifying, and sizing airway and artery trees, we computed the AARs by pairing airway lumens to their closest artery and the percentage of AAR greater than 1 (ai-%AAR>1), with higher values indicating a higher burden of bronchiectasis at CT. Exacerbations were prospectively assessed biannually. We used multivariable, zero-inflation regression to assess the association between ai-%AAR>1 and future exacerbations.
Results: Overall, after adjusting for age, sex, race, body mass index, pack-years, current smoking, oxygen saturation, FEV1, CT measures of emphysema and wall thickness, and CT scanner model/make, ai-%AAR>1 (quartile 4 vs. 1) was associated with an increased risk of future exacerbations (Relative Risk (RR) [95% Confidence Interval (CI)] 1.08 [1.02 – 1.15]). When clinical manifestations (cough, phlegm, dyspnea, exacerbation history) were added to ai-%AAR>1 to define bronchiectasis, the RR [95% CI] was 1.44 [1.37 – 1.50]. Among participants with COPD, corresponding RRs [95% CI] were: 1.10 [1.02 – 1.18] and 1.39 [1.32 – 1.46].
Conclusion: In smokers with COPD, ai-based CT measures of bronchiectasis are associated with an increased future exacerbation risk.