Abstract
Herein, we evaluated the effectiveness of assessing “wetness” of sound extracted from ordinary Japanese utterances to detect dysphagia. Forty-five patients experiencing dysphagic symptoms after undergoing treatment for head and neck cancer were included as participants. We extracted “a” sound from the Japanese word for “month” （“gatsu”）, which was recorded before videofluoroscopic swallowing study （VFSS）. Forty-five “a” sounds were used for perceptual assessments. The recorded VFSS images were classified into the following groups based on the penetration-aspiration scale （PAS）: acceptable （PAS 1 or 2） and PA （PAS 3–8）. The speech samples were evaluated by 16 clinicians associated with dysphagia management. The degree of “wetness” of speech samples was perceptually assessed using a 5-point scale blinded to VFSS data. Average sensitivity and specificity of the first and the second perceptual evaluations were 69.8％ and 32.8％, respectively. Intraevaluator reliability between the first and second perceptual evaluations was 0.82, and interevaluator reliabilities of the first and second perceptual evaluations were 0.90 and 0.91, respectively. Our results suggested that solely relying on abnormal vocal quality is not useful. However, perceptual evaluation of continuous utterances appears promising as a valuable screening method for dysphagia.
