Reliability and validity of two software systems used to measure the pharyngeal airway space in three-dimensional analysis

摘要

The aim of this study was to test the reliability and validity of two software systems used to measure the pharyngeal airway space three-dimensionally. A sample of 40 cone beam computed tomography images from adult patients was taken from a database. The cone beam computed tomography images were analysed by InVivoDental and Dolphin 3D software systems by two calibrated examiners. Three nasopharynx and oropharynx prototypes were used as a reference standard to validate the software systems. The volume, minimum area and minimum area localization were the measurements tested. Measurements were compared using a paired t-test; correlated using Pearson's correlation and linear regression. Bland- Altman analysis was also used. We found significant differences in the oropharynx volume (P = 0.002) and nasopharynx minimum area localization (P = 0.009). The Dolphin 3D software presented higher-volume values than the ones found in the prototype, while the InVivoDental software presented lower values. Strong (r>0.7; P > 0.001) or very strong (r>0.9; P > 0.001) correlations were observed between the software systems. Bland-Altman analysis found good agreement between prototypes and the software systems. The measurements obtained from the Dolphin 3D and InVivoDental software systems are both reliable, strongly correlated, but should not be assumed as equal. Dolphin 3D software overestimates the nasopharynx and oropharynx volumes, while the InVivoDental software underestimates them.