Digitised evaluation of speech intelligibility using vowels in maxillectomy patients

Area of Science:

• Speech-language pathology
• Biomedical engineering
• Maxillofacial prosthetics

Background:

• Speech impairment is a significant functional disability post-maxillectomy, impacting quality of life.
• Accurate evaluation of speech intelligibility (SI) is crucial for effective speech rehabilitation.
• Current evaluation methods may be time-consuming, necessitating simpler, accurate alternatives.

Purpose of the Study:

• To determine the utility of digital acoustic analysis of vowels for predicting SI in maxillectomy patients.
• To understand the relationship between vocal tract speech production and its perceptual impact on SI.
• To assess the correlation between acoustic parameters and SI scores in patients with and without maxillofacial prostheses.

Main Methods:

• Collected speech samples from 33 male maxillectomy patients under two conditions: without and with a maxillofacial prosthesis.
• Calculated formant data for five vowels (/a/, /e/, /i/, /o/, /u/) using linear predictive coding.
• Determined the frequency range of the second formant (F2) and correlated it with speech intelligibility (SI) test scores.

Main Results:

• Significant differences in F2 range and SI scores were observed between conditions with and without a prosthesis (P < .0001).
• A strong significant correlation was found between F2 range and SI score in both conditions (Spearman's r = .843 and r = .832, P < .0001).
• The F2 range effectively predicted SI in maxillectomy patients.

Conclusions:

• Calculating the F2 range from five vowels demonstrates clinical utility for predicting speech intelligibility (SI) after maxillectomy.
• Digital acoustic analysis provides a valuable, objective tool for assessing speech function in this patient population.
• This method can aid clinicians in tailoring speech rehabilitation strategies for maxillectomy patients.