The reliability of upper airway landmark identification

Research involving the upper airway has assessed linear measurements of certain structures on the assumption that the landmarks involved can be reliably identified. This study was conducted to determine the reliability of landmark identification for those structures most commonly reported in the obstructive sleep apnoea literature. Three judges were asked to identify specific landmarks on 20 randomly selected radiographs and 10 superior quality radiographs. This was repeated one week later. The results of the analysis of variance (ANOVA) indicated that the majority of the landmarks could be reliably identified, irrespective of the quality of the radiograph. However, the quality of the radiograph did affect identification of the horizontal position of the hyoid bone and the linear measure of posterior airway space although these were not clinically significant. The vertical position of the tip of the soft palate was highly unreliable, irrespective of the quality of the radiograph. This resulted in errors in the measurement of soft palate length. Future airway-related research should consider the potential inaccuracies when attempting to identify these dynamic three-dimensional structures on static two-dimensional images.