A single blinded prospective cohort study on the diagnostic accuracy of laryngeal ultrasound

Paediatric laryngeal pathology is common, with laryngomalacia being the leading cause of infant stridor and vocal cord nodules the most frequent cause of dysphonia. Although awake flexible nasopharyngolaryngoscopy (FNPLS) provides definitive diagnosis, it is invasive and uncomfortable. Laryngeal ultrasound (LUS) is a non-invasive, radiation-free, and well-tolerated alternative that allows dynamic assessment in awake children.

To compare the reproducibility of FNPLS and LUS findings and to evaluate patient comfort for each procedure.

A single-blinded prospective cohort study was conducted in a tertiary centre in Klang Valley, Malaysia, involving 32 children aged 0-12 years with stridor or dysphonia. FNPLS was performed by an ENT surgeon, followed by LUS within two weeks by a radiologist blinded to FNPLS findings. Sensitivity and specificity of LUS were calculated using FNPLS as the gold standard. Comfort was assessed using VAS or FLACC scales.

LUS showed the highest sensitivity and specificity for subglottic stenosis, followed by laryngeal mass lesions, vocal cord paresis/paralysis, vocal cord masses, and laryngomalacia. Overall specificity was high with variable sensitivity. LUS was significantly more comfortable than FNPLS (mean difference -5.66, p<0.001).

While FNPLS remains the gold standard, LUS is a valuable, less traumatic adjunct, particularly for subglottic stenosis, laryngeal masses, and follow-up evaluation.