Inter-observer agreement in identifying traction bronchiectasis on computed tomography: its improvement with the use of the additional criteria for chronic fibrosing interstitial pneumonia

Purpose To assess inter-observer variability in identifying traction bronchiectasis on computed tomography (CT) using additional criteria for chronic fibrosing interstitial pneumonia. Methods Seven experts categorized CT image set representing 39 patients into three groups on the basis of the presence of traction bronchiectasis, using a three-point scale: 3-definitely/probably yes; 2-possibly yes; and 1-definitely/probably no. This scale served as a reference standard. The image set included cases of chronic fibrosing interstitial pneumonia, non-interstitial lung disease, and difficult-to-determine cases. Forty-eight observers similarly assessed the same image set, first according to the Fleischner Society definition, and second with additional criteria, in which traction bronchiectasis was observed exclusively in chronic fibrosing interstitial pneumonia. The agreement level between the reference standard and each observer's evaluation in each session was calculated using weighted kappa values which were compared between the two sessions using a paired t test. Results The mean weighted kappa value for all observers was significantly higher in the second reading session (mean 0.75) than in the first reading session (mean 0.62) (p < 0.001). Conclusion Inter-observer agreement in identifying traction bronchiectasis improves when using the additional criteria which specify chronic fibrosing interstitial pneumonia as the underlying disease.