Agreement Between 2D and 3D Echocardiography in Measuring Dimensions of the Patent Ductus Arteriosus in Infants

The patent ductus arteriosus (PDA) is a common cardiac lesion in neonates, which may require intervention for patency, or closure in the neonatal period. Two-dimensional echocardiography (2DE) is standard for PDA imaging. Three-dimensional echocardiography (3DE) is increasingly used for assessing complex anatomy and pre-procedural planning; however, there are limited data on the value and accuracy of 3DE of the PDA in infants. We aimed to determine the degree of agreement between 2 and 3DE in measuring common dimensions of the PDA in infants. Infants < 1-year-old, > 32 weeks corrected gestational age (CGA), and > 1 kg with a known PDA were enrolled prospectively for imaging by 2DE and 3DE. Images were collected at the parasternal short axis (PSAX) and suprasternal notch (SSN). Dimensions were measured at the pulmonic (PA) and aortic (Ao) end. Interclass correlation (ICC), Bland-Altman (BA), and coefficient of variability (COV) assessed interobserver variability. 2DE to 3DE association was assessed by BA. Twenty-nine subjects were enrolled and had complete data sets, median CGA 39.4 weeks (IQR 38.3-41.7), median weight 3.250 kg (IQR 2.870-4.295). ICC for all 2D and 3D images in PSAX and SSN views was strong with narrow limits of agreement (LOA). BA showed low bias and generally narrow LOA. Each observer's 2DE- 3DE comparison yielded low bias and narrow LOA. There was generally no statistically significant difference in PDA size when 2D and 3D images were compared for each observer. Our data suggest 2DE and 3DE have strong agreement and paired 2DE- 3DE images show low bias and limited variability for common PDA measures. Previous work in this area has focused on larger patients, and this data begins to build a foundation for use of 3DE for anatomic assessment and possible interventional planning in smaller patients with PDAs. Further study should aim to compare 3DE agreement with CT or angiography, and at variable subject size, ductal size, and ductal type.