3, University of California San Diego, La Jolla, California, United States
1, University of California San Diego, San Diego, California, United States
5, The University of Texas MD Anderson Cancer Center, Houston, Texas, United States
4, The University of Texas MD Anderson Cancer Center, Houston, Texas, United States
Wearable sensors offer the ability to monitor the rehabilitation of patients outside of the clinic. For patients with cancer of the head and neck who have undergone radiation therapy, diminished swallowing activity and treatment effects can result in disuse atrophy and fibrosis of the swallowing muscles in up to 39% of these patients. This condition causes dysphagia and reduced swallowing functionality. There is a need for earlier detection of radiation-associated dysphagia such that subtle changes in swallowing muscle function can be detected well before irreversible damage has occurred. This presentation describes a highly sensitive, flexible strain sensor comprising palladium nanoislands on single-layer graphene. These sensors were placed on the submental region of the neck in a cohort of 14 cancer-free head and neck cancer patients status post radiation therapy with different levels of swallowing function: from non-dysphagic to severely dysphagic. The patch-like devices successfully detect differences in the consistencies of food boluses when swallowed (i.e. water, yogurt, or cracker), along with differences between dysphagic and non-dysphagic swallows. When electrical activity from surface electromyography (sEMG) is obtained simultaneously with the strain data, it is also possible to differentiate swallowing vs. non-swallowing events (i.e. head turns or coughing). The major features in the plots of resistance (strain sensors) and electrical activity (sEMG) vs. time are correlated to specific events during the course of swallowing a barium paste as recorded by video X-ray fluoroscopy (the current standard of care). Finally, we developed a machine learning algorithm to automate the identification of bolus type being swallowed by a healthy subject (Male, 24 years old) with an accuracy of 86.4%. Moreover, the algorithm was also able to discriminate between swallows of the same bolus from either the healthy subject or a dysphagic patient with an accuracy of 94.7%. Taken together, these results may lead to non-invasive and home-based systems for monitoring of swallowing function and improved quality of life.