Serial lung ultrasound scoring for predicting invasive mechanical ventilation in preterm neonates: a prospective diagnostic accuracy study.

Continuous Positive Airway Pressure (CPAP) is the first-line respiratory support for preterm neonates with respiratory distress, but 15-30% fail and require invasive mechanical ventilation (IMV). Lung ultrasound (LUS) has emerged as a radiation-free, point-of-care imaging modality with demonstrated superiority over chest radiography for diagnosing neonatal respiratory conditions, whose serial performance for predicting IMV across a broad preterm cohort is not well characterised. Objective: To evaluate the diagnostic accuracy of serial LUS scores on Days 1-3 for predicting IMV within 7 days in preterm neonates (< 37 weeks), and to identify independent predictors of IMV.  Methods: This prospective diagnostic accuracy cohort study was conducted in a Level III NICU between November 2022 and November 2024. 207 preterm neonates with respiratory distress underwent standardised 6-zone, 18-point lung ultrasound (LUS) scoring on Days 1, 2, and 3. ROC analysis determined optimal LUS score cut-offs for predicting IMV, while multivariable logistic regression identified independent predictors after multicollinearity assessment, model performance was evaluated using goodness-of-fit testing. Results: Among 207 preterm neonates (gestational age 32.5±2.5 weeks; birth weight 1779±626 g), 28 (13.5%) required IMV by Day 7. Day-1 LUS ≥4 predicted IMV with AUC 0.72 (95% CI 0.59-0.84), sensitivity 64%, specificity 78%, NPV 93%; Day-2 LUS ≥2 improved specificity (87%) and NPV (97%); Day-3 LUS showed limited accuracy (AUC 0.61). Surfactant use, shock, and SNAPPE-II score independently predicted the requirement of IMV within the first 7days of life.

conclusionsSerial lung ultrasound scoring demonstrated good clinical utility, with a high negative predictive value (NPV) of the Day-1 LUS score (cut-off ≥ 4) for identifying preterm neonates unlikely to require invasive mechanical ventilation within the first 7 days of life. Integration with clinical scoring may optimize early respiratory triaging in resource-limited NICUs. Multicentre validation against newer consensus frameworks is needed before routine adoption. WHAT IS KNOWN: • LUS is an accurate, radiation-free bedside tool for diagnosing neonatal respiratory disorders, predicting surfactantrequirement in preterm infants. • Most studies have used single-time-point LUS assessments and focused on a single outcome. WHAT IS NEW: •A Day-1 LUS score ≥4 identifi ed infants at low risk of invasive mechanical ventilation (NPV 93%). • Surfactant requirement (OR 28.60), shock (OR 6.18), and SNAPPE-II (OR 1.06 per point) independently predictedIMV.