Abstract
Ultrasound imaging shows promise for bedside tracking of oropharyngeal swallowing structures, but has not been validated against videofluoroscopy. The hyoid bone is readily visible on both ultrasound and fluoroscopy and has multiple muscular attachments involved in execution of swallowing movements. This study aims to validate ultrasound measurements of hyoid motion by quantitatively comparing simultaneously recorded ultrasound and videofluoroscopy data. Participants (N = 23) were recruited from standard-of-care videofluoroscopy studies during which an ultrasound array was positioned submentally to acquire simultaneous B-mode ultrasound imaging. Kinematic analyses of hyoid motion in both ultrasound and videofluoroscopic image data from 56 swallows were analyzed to validate the accuracy of ultrasound measurements. Methods for correction of transducer movement included static angle correction, based on the initial transducer angle, and dynamic translation/rotation, based on the time-dependent transducer position. Hyoid motion variables were compared using intra-class correlation coefficients, root-mean-square error, t tests, and F tests, and included time-dependent trajectories, maximum excursion, and peak velocity. Ultrasound measurements were corrected using dynamic translation/rotation closely agreed with videofluoroscopy (ICC = 0.892 for hyoid trajectories; mean ICC = 0.715 for summary statistics). A substantial portion of hyoid position measurement error was due to transducer motion (estimated as 0.37 cm). These results indicate that ultrasound imaging, when incorporating appropriate correction for transducer motion, can accurately track motion of the hyoid bone and serve as a useful technique for measuring swallowing movements.