The feasibility of establishing agreement between laboratories for measures of oropharyngeal structural movements.

Kinematic analysis, also commonly referred to as biomechanical analysis, of the swallow is used to measure movement of oropharyngeal structures over time. Two laboratory directors who have used kinematic analysis in their research collaborated to determine the feasibility of establishing agreement between two separate laboratories on measures of structural movements of the swallow. This report describes the process that was followed toward the goal of establishing measurement agreement. Under the direction of the laboratory directors, one research technician from each laboratory participated in a process that included initial meetings, training sessions, and pre- and post-training evaluation of reproducibility.

Because agreement on initial measures of structural movement demonstrated weak correlation on some measures, the research technicians trained together for approximately 6 hours. After training, statistical analyses indicated that (a) most Pearson correlations for measures of structural movements were greater than 0.80 and were highly statistically significant; (b) most percentages of absolute deviation were under 25%; and (c) most concordance coefficients were above .70. These statistics indicate that the two laboratories were able to increase their level of agreement in measuring selected structural movements of the swallow after a brief amount of training.

Factors affecting measurement agreement include image quality, establishment of rules for measuring, and the opportunity for regular discussions among research assistants and investigators from both laboratories.

INTRODUCTION

The most commonly used technique to observe the swallow, diagnose dysphagia, and develop a treatment plan for patients with impaired swallow function is the modified barium swallow (MBS) procedure with videofluorography (VFG). (1-3) Investigators in the area of deglutition and dysphagia also use this clinical tool in their research by measuring various aspects of the swallow from the recorded fluoroscopic image. A common measurement protocol is assessment of temporal measures of the movement of the bolus through the oral cavity, pharynx, and cervical esophagus as well as duration of specific events in the oropharyngeal swallow. Examples of temporal measures include durations of oral transit time, pharyngeal transit time, esophageal transit time, laryngeal closure, cricopharyngeal opening, and laryngeal elevation, as well as the temporal organization of these events. (4-7) These measures have become the foundation for the assessment of different cohorts, such as head and neck cancer patients undergoing different types of primary treatment, (8-11) stroke patients (12) and non-patient volunteers. (13), (14) These measures have been used to evaluate the effects of intra-oral prostheses (15) and bolus volume and age (13) on swallowing function.

Kinematic or biomechanical analysis of structural movements critical to swallow has been developed as a method for characterizing swallow function. (17) Kinematic analysis is used to mark the movements of specific oropharyngeal structures over time or at specific points in time from video-fluoroscopic studies. Although measurement algorithms vary, the general procedures for kinematic analysis are consistent across programs. Each video image of interest is digitized; after digitizing, points of interest can be marked on each video image. For most applications, regardless of which software package is used, the researcher will mark an anchor point (a stable, non-moving point in the image), an angle measuring the subject's head tilt relative to true vertical, and a reference distance. These elements are marked on each video image and used in conjunction to control for a subject's head and body movement in the vertical, A-P, and horizontal planes as well as to measure absolute distances of structural movement during the swallow. On each marked video image, the anchor point is adjusted to become the origin of the coordinate system by subtracting its actual x and y pixel values from each point in the digitized image. In this manner, the anchor point on each image always has the adjusted coordinates of (0,0), regardless of how much a subject might move in any plane. The subject's postural angle is also used to virtually rotate each individual digitized image so that the subject's head tilt is mathematically in line with true vertical. This adjustment simplifies calculation of anterior and vertical movement of structures and assures that structural movement is calculated relative to the vertebral column. Finally the reference distance is used to scale the marked points to actual size or to some relative value such as vertebral body units. Because the reference distance is marked on each digitized image, it also adjusts for any movement the subject might make during the study toward or away from the fluoroscope and corrects for differences in fluoroscopic magnification during the swallow. Structures of interest are then marked and calculated relative to the anchor point, angle of head tilt, and reference distance. Figure 1 illustrates an example of a marked image.

[FIGURE 1 OMITTED]