Abstract
OBJECTIVE: To determine the sensitivity and specificity of lung ultrasound in diagnosing respiratory pathologies in children on respiratory support in the pediatric intensive care unit, compared to chest X-ray and clinical diagnosis. METHODS: A cross-sectional study was conducted on children aged 1 month to 18 years admitted to the pediatric intensive care unit and requiring respiratory support from June 2018 to February 2019. Lung ultrasound was performed within 24 hours of chest X-ray by a trained sonographer using standardized protocols. Lung ultrasound and chest X-ray were interpreted independently by blinded physicians. A Receiver Operating Characteristic curve was generated to assess lung ultrasound diagnostic performance using chest X-ray as the gold standard. RESULTS: A total of 220 lung ultrasounds were performed on 117 patients, with 195 (88.6%) examinations completed. Lung ultrasound and chest X-ray were reported normal in 24 (10.9%) and 21 (9.5%) studies, respectively, with no pneumothorax detected. Overall, lung ultrasound had a sensitivity of 89.95% and specificity of 19.05% compared to chest X-ray. Sensitivity and specificity for pneumonia and pediatric acute respiratory distress syndrome were 62.8% and 44.8%, and 50% and 96%, respectively. Using clinical diagnosis as reference, sensitivity and specificity for pneumonia and pediatric acute respiratory distress syndrome were 59.8% and 54.3%, and 72.7% and 95.7%, respectively. Agreement between chest X-ray and lung ultrasound was poor (k = 0.085), though concordance among lung ultrasound providers was high (k = 0.869). Agreement for lung ultrasound and pediatric acute respiratory distress syndrome was highest (k = 0.632) compared with clinical diagnosis. Receiver Operating Curve analysis showed that lung ultrasound showed poor diagnostic accuracy compared to chest X-ray (AUC 0.54). CONCLUSION: Lung ultrasound is feasible in low-resource pediatric intensive care unit settings but shows limited diagnostic accuracy compared to chest X-ray.