ICHNO-ECHNO 2022 - Abstract Book

S104

ICHNO-ECHNO 2022

difference was significant (p<0.05). Mean submandibular gland dose did not correlate with composite xerostomia scores but correlated with the unstimulated xerostomia scores at 3 months (p<0.05, r=0.7) Conclusion Xerostomia worsened significantly at 3 months and improved at 6 months post treatment. Scores assessing xerostomia in the unstimulated setting correlated with mean submandibular gland dose. Almost all patients in the locally advanced setting had doses to the submandibular gland exceeding recommended constraints. There was significant correlation between submandibular gland dose and dose to the parotids and oral cavity. Location of the primary has an important role in determining the dose to the submandibular gland. References 1. Eisbruch A, Kim HM, Terrell JE, Marsh LH, Dawson LA, Ship JA. Xerostomia and its predictors following parotid-sparing irradiation of head-and-neck cancer. International Journal of Radiation Oncology* Biology* Physics. 2001

PO-0165 Title: Acoustic analysis of voluntary and reflexive coughs in healthy subjects

S. Mootassim-Billah 1

1 Jules Bordet Institute, Radiation-Oncology, Brussels, Belgium

Purpose or Objective Late radiation-associated dysphagia (RAD) is defined as impaired swallowing efficiency/safety following chemoradiotherapy in head and neck cancer patients. In a RAD framework, the risk of impaired coughing may lead to lung aspiration and fatal lung infection. Although cough efficacy is a predictor of aspiration, cough investigation is minimal in patients with RAD. Acoustically, a cough sound is a superposition of turbulence noise – presumably due to the air jet passing through the glottis – and lower frequency noise – presumably generated by the airflow modulated by the vibrations of the vocal folds, false vocal cords and aryepiglottic folds. Because cough is a transient signal, existing software for speech analysis are not appropriate. The goal of our project is to develop an assessment method using acoustic features related to voluntary and reflexive coughs as biomarkers of aspiration in patients with RAD. As our methods are new, this study first focuses on healthy subjects in order to obtain reference values. Materials and Methods Forty healthy subjects produced voluntary and (induced) reflexive coughs, recorded using a throat microphone. Cough samples were analyzed with a software under development. Automatic final segmentation enables to measure the cough duration. A first method of analysis includes temporal features: the amplitude contour, the sample entropy and the kurtosis. These features report respectively the strength, the unpredictability (turbulence noise due to the air jet) and the impulsive quality (burst) of the signal. A second method of analysis consists of a spectral decomposition of the relative cough signal energy into several frequency bands (0-400Hz, 400-800Hz, 800-1600Hz, 1600Hz-3200Hz, >3200Hz). Results Results show an average duration of 0.3 sec for voluntary and reflexive coughs. The temporal analysis (Table 1) reports that reflexive coughs exhibit higher feature values for the curvature of the amplitude contour and the average and curvature of the kurtosis compared to voluntary coughs. This suggests overall higher impulsivity and higher contrast of impulsivity between onset, sustain and offset in reflexive coughs. The average and curvature of the sample entropy are also higher in reflexive coughs, suggesting overall higher flow noise and higher contrast of flow noise between onset, sustain and offset. Overall higher flow noise is confirmed by the larger weighted frequency and by the higher spectral features (Table 2) in frequency bands >800 Hz. Voluntary and reflexive coughs show comparable energy in frequencies <800Hz, suggesting comparably intense mechanical vibrations of the glottal structures.