Extension and Application of High-Speed Digital Imaging Analysis Via Spatiotemporal Correlation and Eigenmode Analysis of Vocal Fold Vibration Before and After Polyp Excision

Abstract

Objectives/Hypothesis: To evaluate the spatiotemporal correlation of vocal fold vibration using eigenmode analysis before and after polyp removal and explore the potential clinical relevance of spatiotemporal analysis of correlation length and entropy as quantitative voice parameters. We hypothesized that increased order in the vibrating signal after surgical intervention would decrease the eigenmode-based entropy and increase correlation length. Study Design: Prospective case series. Methods: Forty subjects (23 males, 17 females) with unilateral (n = 24) or bilateral (n = 16) polyps underwent polyp removal. High-speed videoendoscopy was performed preoperatively and 2 weeks postoperatively. Spatiotemporal analysis was performed to determine entropy, quantification of signal disorder, correlation length, size, and spatially ordered structure of vocal fold vibration in comparison to full spatial consistency. The signal analyzed consists of the vibratory pattern in space and time derived from the high-speed video glottal area contour. Results: Entropy decreased (Z = −3.871, P < .001) and correlation length increased ( t = −8.913, P < .001) following polyp excision. The intraclass correlation coefficients (ICC) for correlation length and entropy were 0.84 and 0.93. Conclusion: Correlation length and entropy are sensitive to mass lesions. These parameters could potentially be used to augment subjective visualization after polyp excision when evaluating procedural efficacy.