Yuan-Jia Hu, Jia Song, Sen-Nian Zheng, Lu Xing, Yu-Jie Feng, Sheng-Tao Wu, Shu-Yi Zeng, Ying-Shan Tan, Qiu-Tong Chen, Pei-Yun Zhuang
HSV provides multidimensional, objective, and quantitative metrics for LD assessment, reducing reliance on subjective evaluation. Future studies should employ multicenter designs with larger cohorts to standardize parameter definitions and further validate the reliability of AI-assisted analytical approaches, ultimately promoting more objective and standardized diagnosis of LD.
OBJECTIVE: To review the clinical application and quantitative parameters of high-speed videoendoscopy (HSV) in evaluating vocal fold kinematics in patients with laryngeal dystonia (LD), to summarize the limitations of existing studies, and to provide a reference for future clinical standardization.
METHOD: A comprehensive search of eight databases, including PubMed, Embase, and the Web of Science Core Collection (WoSCC), was conducted from inception to December 2025. Three independent investigators performed the literature screening and data extraction. A scoping review and visual analysis were subsequently conducted on the included studies.
RESULTS: Across the included studies, HSV captured several LD-related abnormalities, including oscillatory breaks, micromotions, motion irregularities, supraglottic compression, altered open quotient (OQ), speed quotient (SQ), glottal area change index (GACI), vocal vibration opening onset position (VVOOP), and prolonged or more variable glottal attack time (GAT) and glottal offset time (GOT). These findings were organized into five parameter domains, covering abnormal movements and configuration, micro-vibratory dynamics and kinematics, onset/offset temporal dynamics, image-field usability, and automated analysis metrics. Recent studies further incorporated artificial intelligence (AI)-assisted approaches, including automated obstruction detection and glottal segmentation.
CONCLUSIONS: HSV provides multidimensional, objective, and quantitative metrics for LD assessment, reducing reliance on subjective evaluation. Future studies should employ multicenter designs with larger cohorts to standardize parameter definitions and further validate the reliability of AI-assisted analytical approaches, ultimately promoting more objective and standardized diagnosis of LD.