Automated quantitative analysis of vocal fold vibration using two-dimensional scanning videokymography after transoral laser microsurgery

Abstract

The purpose of this study was to develop a quantitative automated analysis system for two-dimensional videokymography (2D VKG) using the threshold segmentation method, and to analyze the vocal fold vibration characteristics of vocally healthy subjects and patients who underwent CO2 transoral laser microsurgery using this system, and to examine in accuracy differences between automatic and automated analysis methods.Methods: 29 male patients who underwent CO2 transoral laser microsurgery and 10 vocally healthy male participated in the study. Quantitative analysis such as opening quotient (OQ), phase symmetry index (PSI), amplitude symmetry index (ASI) of 2D VKG images was performed using an automated analysis program in which glottal area extraction by the threshold segmentation method and edge detection using a manual plotting technique were mixed.Results: Automatic analysis enabled accurate quantitative analysis in 76.9% (30/39) of the total image. The automated analysis corrected by the manual plotting technique on the sample showing gross error showed higher accuracy than the automatic analysis. The results of automated and automatic analysis were not statistically different in ASI, but there were differences in OQ and PSI. In addition, vocally healthy group and patients group showed statistical significance in all parameters (OQ, ASI, PSI). Conclusions:The 2D VKG, which can evaluate the vibration of the entire vocal fold tissue in real time, can be usefully utilized as a laryngeal imaging technique that visualizes the structure and function of the vocal fold. Quantitative analysis using automated and automatic analysis can increase the clinical usability of 2D VKG.