A cross-sectional protocol for experimental tongue high-density surface electromyography to detect and classify radiation-associated hypoglossal neuropathy.
Hypoglossal neuropathy is the most common lower cranial neuropathy detected as a delayed sequelae of Human Papillomavirus (HPV) -driven oropharyngeal cancer (OPC). Needle electromyography (EMG) is the gold standard for electrodiagnostic testing, but it is invasive and relies on subjective interpretation of the EMG signal. This study explores the potential of non-invasive high-density surface electromyography (HDSEMG) to detect and quantify hypoglossal neuropathy in OPC survivors.
In an exploratory study, examine the feasibility of HDSEMG for rapid, non-invasive screening of hypoglossal nerve (CN XII) function and estimate the prevalence of hypoglossal neuropathy before and after oropharyngeal radiotherapy, and associate with patient-reported and clinician-graded functional outcomes. Machine learning performance will be measured through sensitivity, specificity, and F1 score, with a target area under the curve > 0.7 based on literature-reported EMG sensitivity and specificity.
This protocol will recruit patients aged ≥ 18 years who receive radiation therapy for OPC at MD Anderson Cancer Center (MDACC) between 2024-2025 and consent to experimental HDSEMG testing. Sanchez Research Lab (The University of Utah, Salt Lake City, UT) will perform data analysis. Clinical data-including electrical impedance measurement (EIM), patient-reported outcomes, dysphagia grading, tongue functions, fibrosis grading, and needle EMG-will be collected from n = 36 patients. Features extracted from HDSEMG will be correlated with other clinical outcomes and used to train a machine learning classifier to quantify the severity of hypoglossal neuropathy.
In an exploratory study, examine the feasibility of HDSEMG for rapid, non-invasive screening of hypoglossal nerve (CN XII) function and estimate the prevalence of hypoglossal neuropathy before and after oropharyngeal radiotherapy, and associate with patient-reported and clinician-graded functional outcomes. Machine learning performance will be measured through sensitivity, specificity, and F1 score, with a target area under the curve > 0.7 based on literature-reported EMG sensitivity and specificity.
This protocol will recruit patients aged ≥ 18 years who receive radiation therapy for OPC at MD Anderson Cancer Center (MDACC) between 2024-2025 and consent to experimental HDSEMG testing. Sanchez Research Lab (The University of Utah, Salt Lake City, UT) will perform data analysis. Clinical data-including electrical impedance measurement (EIM), patient-reported outcomes, dysphagia grading, tongue functions, fibrosis grading, and needle EMG-will be collected from n = 36 patients. Features extracted from HDSEMG will be correlated with other clinical outcomes and used to train a machine learning classifier to quantify the severity of hypoglossal neuropathy.
Authors
Hansen Hansen, Woodman Woodman, Peterson Peterson, Buoy Buoy, Tang Tang, Mao Mao, Moreno Moreno, Lai Lai, Fuller Fuller, Barbon Barbon, McMillan McMillan, Anderson Anderson, Hutcheson Hutcheson, Sanchez Sanchez,
View on Pubmed