Quantifying lung aeration in neonatal lambs at birth using lung ultrasound

Abstract

BackgroundLung ultrasound (LUS) is a safe and non-invasive tool that can potentially assess regional lung aeration in newborn infants and reduce the need for X-ray imaging. LUS produces images with characteristic artifacts caused by the presence of air in the lung, but it is unknown if LUS can accurately detect changes in lung air volumes after birth. This study compared LUS images with lung volume measurements from high-resolution computed tomography (CT) scans to determine if LUS can accurately provide relative measures of lung aeration.MethodsDeceased near-term newborn lambs (139 days gestation, term ∼148 days) were intubated and the chest imaged using LUS (bilaterally) and phase contrast x-ray CT scans at increasing static airway pressures (0–50 cmH2O). CT scans were analyzed to calculate regional air volumes and correlated with measures from LUS images. These measures included (i) LUS grade; (ii) brightness (mean and coefficient of variation); and (iii) area under the Fourier power spectra within defined frequency ranges.ResultsAll LUS image analysis techniques correlated strongly with air volumes measured by CT (p < 0.01). When imaging statistics were combined in a multivariate linear regression model, LUS predicted the proportion of air in the underlying lung with moderate accuracy (95% prediction interval ± 22.15%, r2 = 0.71).ConclusionLUS can provide relative measures of lung aeration after birth in neonatal lambs. Future studies are needed to determine if LUS can also provide a simple means to assess air volumes and individualize aeration strategies for critically ill newborns in real time.

Point-of-Care Ultrasound versus Chest X-Ray for Determining Lung Expansion Based on Rib Count in High-Frequency Oscillatory Ventilation

“…There are a few experimental data about absolute lung volume during conventional ventilation or HFOV with computed tomography, but it is quite challenging to measure lung capacity, especially during HFOV in critically ill neonates [9, 10]. CXR, which is mostly performed in the anteroposterior supine position, is sufficient for the evaluation of the chest wall, heart, lungs, diaphragm, catheter, and endotracheal tubes and can also be used for approximation.…”

Assessing lung aeration using ultrasound after birth in near-term lambs at risk of respiratory distress

“…A random sample of 92 images was graded by a second blinded researcher (DB), and the inter-rater reliability was calculated and reported as a percentage. LUS coefficient of variation (CoV) of pixel intensity: The estimated proportion of air (EPA) in the lungs was calculated using a recently described technique ( 14 ). This technique involves manually defining rectangular regions of interest in each ultrasound recording, defined as any area of lung below the pleural line and between acoustic shadows formed by the ribs ( Figure 1 ).…”

“…Frames coinciding with large amounts of movement were excluded from this analysis. The mean CoV in these regions was calculated by dividing the standard deviation of pixel intensity by the mean pixel intensity, and a simple linear regression model was used to determine the EPA from the CoV as previously described ( 14 ). The EPA ranges from 0 (lung is completely liquid filled) to 1 (lung is completely air-filled).…”

“…LUS images have previously been scored based on the presence or absence of different image artifacts, which provide a rough indication of the degree of lung aeration ( Figure 1 ). These imaging artifacts have been well characterised and correlate well with the amount of air in the lungs of newborn lambs ( 14 ). Indeed, LUS can describe changes during the transition to air-breathing at birth in healthy term infants ( 12 , 13 , 15 , 16 ), term infants with respiratory distress ( 12 ), and preterm infants ( 12 , 17 ).…”

“…( B ) Example lung ultrasound image with three regions of interest (ROI; area below the pleural line, not including acoustic shadows) selected. The coefficient of variation of pixel intensity (LUS CoV) was calculated from these ROIs, and used to estimate the proportion of air in the lungs as previously described ( 14 ).…”