Electroglottogram approximate entropy: a novel single parameter for objective voice assessment.

Abstract

BACKGROUND: The electroglottogram approximate entropy value is a numerical variable which gives an overall measure of voice quality. It is derived by analysing the complexity of the electroglottogram waveform using regulatory statistics. AIMS: (1) To use electroglottogram approximate entropy to measure voice quality in patients with glottic pathology and in normal subjects, to ascertain whether this parameter can distinguish between pathological and normal voices. (2) To ascertain whether electroglottogram approximate entropy can measure voice change over time within individual subjects. (3) To determine any correlation between electroglottogram approximate entropy and the grade-roughness-breathiness-asthenia-strain scale. METHODS: One hundred and forty-one normal volunteers were recruited to characterise electroglottogram approximate entropy in the normal voice. One hundred and eighty-six patients with glottic squamous cell carcinoma underwent electroglottogram approximate entropy measurement prior to radiotherapy and then three to six months and one year after treatment. Subjects' voices were categorised by a speech therapist using the grade-roughness-breathiness-asthenia-strain scale. RESULTS: The mean electroglottogram approximate entropy of the normal volunteers was 0.302 (range 0.05-0.42). The mean electroglottogram approximate entropy of the glottic squamous cell carcinoma patients was significantly lower prior to treatment, at 0.227 (range 0.001-0.397; p < 0.0005), but improved after radiotherapy to 0.277 at three to six months and 0.282 at one year. Electroglottogram approximate entropy results correlated significantly with grade-roughness-breathiness-asthenia-strain scale results. CONCLUSION: Electroglottogram approximate entropy can be used to assess change in voice quality resulting from glottic morphological abnormality. Electroglottogram approximate entropy values improve as voice quality improves after treatment. Electroglottogram approximate entropy values correlate significantly with grade-roughness-breathiness-asthenia-strain scale results.