Associations Between Acoustic, Kinematic, Self-Reported, and Perceptual Measures of Speech in Individuals Surgically Treated for Oral Cancer

Abstract

Purpose: The purpose of this study was to assess differences between individuals treated for oral squamous cell carcinoma (ITOC) and control speakers on acoustic, kinematic, and perceptual measures of speech. Furthermore, this study aimed to assess the interrelatedness of these speech domains alongside self-reported speech outcomes in order to inform clinically relevant measures of speech in ITOC. Method: Simultaneous acoustic and kinematic data (via electromagnetic articu-lography sensors on the tongue) were collected from nine ITOC, who received surgical treatment for a tumor located on either the tongue or jaw and eight age-and sex-matched control speakers. All participants were native speakers of Dutch and read the North Wind and the Sun Passage. We calculated the articulatory–acoustic vowel space (AAVS) from the acoustic data and the articulatory–kinematic vowel space (AKVS) from the tongue tip and tongue back sensor data. Inexperienced listeners (n = 35) provided intelligibility and listening effort ratings using a visual analogue scale rating procedure. Self-reported speech problems were assessed using the Speech Handicap Index. Results: Compared to an age-and sex-matched control group, ITOC demonstrated a significantly smaller AAVS and AKVS of the tongue tip and back, as well as lower intelligibility ratings. A correlation analysis of all speech outcome measures within the ITOC group showed that group-wise, the acoustic, perceptual, and self-reported measures were most strongly associated with each other. While acoustic and kinematic measures were not strongly associated with each other on the group level, within-speaker correlations showed stronger acoustic–kinematic associations. Conclusions: This study demonstrates that acoustic, perceptual, and self-reported measures are related and quantify speech problem severity between ITOC, while kinematic measures showed no between-speaker relationships in a systematic way. Acoustic and kinematic measures showed greater within-speaker than between-speaker associations, reflecting speaker-specific compensatory behaviors. Our results underscore the importance of assessing the speech outcomes of ITOC across the acoustic, kinematic, perceptual, and self-reported domains to inform rehabilitation strategies.