Swallowing disorders (dysphagia) are a major complication of numerous diseases and affect the lives of millions of Americans. The """"""""gold standard"""""""" for dysphagia diagnosis is videofluoroscopy. Currently, the biomechanical analysis of videofluoroscopic (VFS) oropharyngeal swallowing is cumbersome, non-standardized, and often lacks objectivity. No software to date provides rapid postprocessing and automated movement analysis for VFS oropharyngeal swallowing. We propose to fill this void by developing such a software package for the swallowing rehabilitation community, especially one that will automatically and accurately track the movement of the hyoid bone - a major biomechanical marker of swallowing. The objective of this SBIR project is for Peak Performance Technologies, Inc. to expand the image processing and movement analysis techniques of its Peak Motus software to produce rapid, accurate on-line measurements of the kinematic and biomechanical parameters of VFS oropharyngeal swallowing. When fully developed, the proposed software will provide swallowing researchers with powerful analysis tools for multi-center clinical trials, benefit educators and students in the academic preparation of future clinicians, and ultimately supply swallowing clinicians with an improved biomechanical foundation for evidence-based dysphagia diagnostics. The Phase I feasibility study will: 1) Develop and test three variations of a fluoroscopic tracking registration algorithm to automatically identify and track movement of the hyoid bone. Specifically, we will: (a) create three synthetic data sets with precise, known hyoid displacements, (b) develop the algorithms for accurate tracking of hyoid movement, (c) apply the hyoid tracking algorithms to each synthetic data set and evaluate tracking accuracy by comparing results to known displacements, and (d) demonstrate the feasibility of applying the tracking algorithms to an actual videofluoroscopic data set. 2) Develop and test algorithms for performing specific temporal and spatial measurements of normal VFS swallowing behaviors based on tracked bolus movement. Phase II wilt have three parts. Part I will create more complex synthetic data sets to further develop accurate automatic tracking of 2-D hyoid motion, integrate all features into a comprehensive swallowing biomechanics analysis package, and conduct validation testing. Part II will stringently test the software on a large number of normal swallows at the NIH. Part III will develop a commercial research product and performance support materials, and conduct extensive beta product testing.
