How accurate is CT morphometry of airway? Phantom and clinical validation study

Abstract

To assess the reliability of CT airway measurement using phantom and to evaluate its reproducibility using clinical chest CT images. The phantom consisted of six tubes with different diameters and wall thicknesses was used. The wall area ratio (%WA) and wall thickness ratio (%WT) were calculated and the difference from the actual value (error ratio) was assessed. In vivo validation was performed with MDCT data of consecutive 10 patients and inter- and intra-rater agreements of the measurement were evaluated. The error ratio of %WA and %WT increased for the phantom tube with a 1mm or thinner wall thickness. The FOV size has an influence on the airway measurement especially for the tube with 1mm wall thickness. Inter-rater reliabilities between two observers for %WA and %WT were excellent and good intra-class correlation coefficient for %WA and %WT were obtained as 0.825 and 0.811, respectively. Intra-rater reliabilities of measurement also showed good intra-class correlation coefficient for %WA and %WT as 0.822 and 0.800, respectively. Although the phantom study showed there is a tendency to overestimate the wall area of an airway with a thickness 1mm or less, the software has a potential to be used in assessing longitudinal observation of the same patient, as well as, comparison among different patients due to good intra- and inter-rater reliability.