Clinical Utility of Acoustic Voice Parameters and Patient Demographic Variables in Identifying Vocal Fold Pathology...

To evaluate the ability of conventional acoustic voice measures and cepstral analysis to distinguish healthy voices from vocal fold pathology and to determine their utility in differentiating benign from malignant vocal fold lesions.

Retrospective observational study.

A total of 177 participants were included: 72 healthy controls, 47 patients with benign vocal fold lesions, and 58 patients with histopathologically confirmed laryngeal squamous cell carcinoma. Acoustic analysis was performed using Praat software. Fundamental frequency (F0), jitter, shimmer, harmonics-to-noise ratio (HNR), and cepstral peak prominence-smoothed (CPPS) were extracted from a standardized 3-second sustained vowel segment. Group comparisons, receiver operating characteristic (ROC) analyses, and multivariable logistic regression were performed.

Significant group differences were observed for jitter, shimmer, HNR, and CPPS (all P < 0.001), whereas F0 did not differ significantly among groups (P = 0.276). Both benign and malignant lesion groups demonstrated significantly higher jitter and shimmer values and significantly lower HNR and CPPS values than healthy controls. In direct comparisons between benign and malignant lesions, only HNR remained significantly different (P = 0.012). For differentiating healthy and pathologic voices, HNR demonstrated the highest diagnostic performance (AUC = 0.849, 95% CI: 0.790-0.908), followed by CPPS (AUC = 0.811), shimmer (AUC = 0.801), and jitter (AUC = 0.779). In contrast, discrimination between benign and malignant lesions was limited, with HNR achieving the highest AUC (0.643). In multivariable analysis, age, sex, and smoking status were independently associated with malignancy, whereas HNR was not.

Acoustic voice measures, particularly HNR and CPPS, demonstrate good performance for identifying vocal fold pathology but have limited utility for distinguishing benign from malignant lesions. These findings suggest that acoustic analysis may serve as a useful screening tool for detecting underlying laryngeal pathology and supporting referral for advanced laryngeal evaluation rather than functioning as a lesion-specific diagnostic marker.