Feasibility and accuracy of laryngeal ultrasound for the assessment of vocal cord mobility in children

Objective: Transcutaneous laryngeal ultrasonography has been utilized to document vocal cord mobility in the adult perioperative thyroidectomy setting with variable success. The aims of this study were to assess the feasibility of using transcutaneous ultrasound in vocal cord visualization in children, to determine the accuracy in detecting vocal cord immobility compare to flexible laryngoscopy, and to identify any barriers that may affect its utility. Methods: This is a prospective blinded-assessor study. Transcutaneous laryngeal ultrasound and flexible laryngoscopy were performed in two pediatric settings: perioperative thyroid surgery and inpatient consultation for airway, voice, or swallowing concerns. The video recordings of ultrasound and laryngoscopy were assessed by two fellowship-trained pediatric otolaryngologists independently. Results: A total of 83 paired laryngoscopy and ultrasound assessments were performed, 39 for perioperative thyroidectomy and 44 for inpatient consultation. The majority had normal vocal cord mobility (65/83, 78%), while 16 had unilateral cord palsy and 2 had bilateral palsy. The vocal cords were successfully visualized on ultrasound in 82/83 (99%) evaluations. Compared to gold standard laryngoscopy, the sensitivity for diagnosing vocal cord palsy on ultrasound was 53-83%, and the specificity was 97-98%. The sensitivity for detecting asymmetric vocal cord movement was 56-88%, and the specificity was 99%. The agreement between ultrasound and laryngoscopy diagnoses were 87-94% (kappa = 0.58-0.83, moderate to near perfect agreement). The intra-rater agreement was 95-100% (kappa = 0.64-1.0), and inter-rater agreement was 90% (kappa = 0.66). Both subjects with bilateral vocal cord palsy had incorrect ultrasound diagnoses by both observers. Conclusion: We were able to utilize transcutaneous laryngeal ultrasound to successfully visualize vocal cord mobility in almost all pediatric patients with a high degree of specificity. The modality has limitations in the assessment of bilateral vocal cord palsy, and its sensitivity was observer-dependent.