Reproducibility of airway wall thickness measurements on CT in a lung cancer screening setting

Background: Airway wall thickness (AWT) is associated with lower lung function in patients with chronic obstructive pulmonary disease (COPD). Longitudinal airway wall thickness measurements may be used to monitor disease progression. However, the variability of AWT measurements is not well known. Aims and objectives: We investigated the reproducibility of AWT on CT and the relation with inspiration level in participants of a lung cancer screening study (NELSON trial). Methods: Participants who underwent follow-up CT after 3 months because of an indeterminate pulmonary nodule were included. AWT was automatically analyzed and expressed as Pi10. Inter-examination reliability was assessed by intra class correlation coefficients (ICC). Limits of agreement were determined with the Bland-Altman approach. Total lung volume on CT was used as a surrogate for inspiration level. Results: In total, 591 men were included with a mean±SD Pi10 at baseline of 2.81±0.80 mm and of 2.81±0.82 mm at follow-up. ICC was 0.70. Limits of agreement for differences in AWT were -1.22 mm to 1.22 mm. Differences in Pi10 showed a strong correlation with variation in level of inspiration (r=0.744, p Conclusions: Measuring AWT on CT with Pi10 in a lung cancer screening setting has good reproducibility, but individual variation can be substantial. An increase in Pi10 of at least 1.22 mm is required for detection of an increase in the extent of airway wall thickness. However, when correcting Pi10 for lung volume changes, limits of agreement narrow and variability improves.