METHODS: Children aged <18-years scheduled for FB and MDCT were recruited. FB and MDCT were undertaken within 30-min to 7-days of each other. Tracheobronchomalacia (mild, moderate, severe, very severe) diagnosed on FB were independently scored by two pediatric pulmonologists; VB was independently scored by two pairs (each pair = pediatric pulmonologist and radiologist), in a blinded manner.

RESULTS: In 53 children (median age = 2.5 years, range 0.8-14.3) evaluated for airway abnormalities, tracheomalacia was detected in 37 (70%) children at FB. Of these, VB detected tracheomalacia in 20 children, with a sensitivity of 54.1% (95%CI 37.1-70.2), specificity = 87.5% (95%CI 60.4-97.8), and positive predictive value = 90.9% (95%CI 69.4-98.4). The agreement between pediatric pulmonologists for diagnosing tracheomalacia by FB was excellent, weighted κ = 0.8 (95%CI 0.64-0.97); but only fair between the pairs of pediatric pulmonologists/radiologists for VB, weighted κ = 0.47 (95%CI 0.23-0.71). There were 42 cases of bronchomalacia detected on FB. VB had a sensitivity = 45.2% (95%CI 30.2-61.2), specificity = 95.5% (95%CI 94.2-96.5), and positive predictive value = 23.2 (95%CI 14.9-34.0) compared to FB in detecting bronchomalacia.