Abstract:Objective: Dysphagia is a significant ischemic stroke complication that can lead to aspiration. Identification of at-risk patients can be logistically difficult and costly. Researchers investigated whether quantitative ultrasound assessment of hyoid bone movement during induced swallowing would predict failure of videofluoroscopy (VFS) or fiberoptic endoscopic evaluation of swallowing (FEES), as determined by a penetration-aspiration scale (PAS) score. Additionally, ability of a machine learning (ML) algorithm to predict PAS success or failure from real-time ultrasound video recordings was assessed. Methods: A prospective, single-blinded, observational pilot study was conducted from June 2019 through March 2020 at a comprehensive stroke center on a convenience sample of patients admitted with diagnosis of acute ischemic stroke undergoing VFS or FEES as part of dysphagia assessment. Researchers performed a midsagittal airway ultrasound during swallowing in patients receiving an objective swallowing assessment by speech language pathologists who were blinded to ultrasound results. Sonologists measured hyoid bone movement, and researchers then constructed an ML algorithm designed for real-time video analysis using a long short-term memory network with an embedded VGG16 convolutional neural network. Results: Videos from 69 patients were obtained with their respective PAS results. In total, 90% of available videos were used for algorithm training. After training, the ML algorithm was challenged with the 10% previously unseen videos and then compared with PAS outcomes. Statistical analysis included logistic regression and correlation matrix testing on human ultrasound measurements. Cohen's κ was calculated to compare deep learning algorithm prediction with PAS results. Measurement of hyoid bone elevation, forward displacement, total displacement and mandible length was unable to predict PAS assessment outcome (p values = 0.36, 0.13, 0.11 and 0.32, respectively). The ML algorithm showed substantial agreement with PAS testing results for predicting test outcome (κ = 0.79; 95% confidence interval: 0.52-1.0) CONCLUSION: Manual ultrasound measurement of hyoid movement during swallowing in stroke patients failed to predict PAS swallowing results. However, an ML algorithm showed substantial agreement with PAS results despite a small data set, which could greatly improve access to dysphagia assessment in the future.

Application of Real-time Submental Ultrasonography to Assess Swallowing

Ultrasound-based Visualization Measurement of Hyoid-Mandibular Motion for Assessing the Efficacy of Acupuncture in Treating Post-Stroke Dysphagia: A Clinical Study

Ultrasonographic Examination and Evaluation of Dysphagia in Stroke Patients

Characterizing Swallows from People with Neurodegenerative Diseases Using High-Resolution Cervical Auscultation Signals and Temporal and Spatial Swallow Kinematic Measurements

Kinematic analysis of hyoid bone and laryngeal displacement in patients with nasopharyngeal carcinoma after radiotherapy

Tracking Hyoid Bone Displacement During Swallowing Without Videofluoroscopy Using Machine Learning of Vibratory Signals

Morphological and functional changes of the geniohyoid muscle in elderly patients after hip fracture surgery: Comparison of ultrasound images with a focus on swallowing function

Neck sensor-supported hyoid bone movement tracking during swallowing

Verification on the Feasibility of A Vision-Based Swallowing Detection Method

Assessing the oral and suprahyoid muscles in healthy adults using muscle ultrasound to inform the swallowing process: a proof-of-concept study

Movement of the Hyoid Bone and Ventriculus Laryngis in Normal Swallowing

COMPARISON OF SUBMENTAL sEMG DURING DRY SWALLOW BETWEEN IRRADIATED HEAD AND NECK CANCER SURVIVORS AND NORMAL INDIVIDUALS

A New Method for Swallowing Detection with Binocular Vision During Radiotherapy

Efficacy survey of swallowing function and quality of life in response to therapeutic intervention following rehabilitation treatment in dysphagic tongue cancer patients.

Bedside Ultrasound to Identify and Predict Severity of Dysphagia Following Ischemic Stroke: Human Versus Artificial Intelligence

Surface Electromyographic Activities of Submental Muscles among Stroke Patients with Dysphagia

Localization of the Submandibular Glands Using High-Frequency Ultrasound

Swallowing Functional Outcomes and Nutritional Status in Head and Neck Cancer Radiotherapy: Longitudinal Study

Chemoradiation to the Submental Muscles Alters Hyoid Movement During Swallowing in a Rat Model

Effect of voice training intervention on swallowing function in patients with head and neck cancer undergoing radiotherapy: A randomized controlled trial

Swallowing in advanced oral cancers: A prospective observational study