Inter- and intra-scanner differences in Doppler flow measurements using three-dimensional printed flow phantom model of the carotid artery in preterm infants

Abstract

This study compared flow measurements with three dimensional (3D) printed flow phantom models of the carotid artery of preterm infants using two commonly available brands of ultrasound scanners (scanner 1 and 2) in the neonatal unit.

3D printed flow phantoms with three different vessel diameters was used for performing continuous and pulsatile flow measurements by 2 observers, blinded to phantom characteristics and each other’s measurements. Agreement between observers were examined using Bland-Altman analysis.

Of the 48 measurements, 62% were pulsatile flow measurements and majority (36(75%)) performed with phantom vessel diameter of 0.244 cm. Bland-Altman analysis for continuous flow measurements (n=18(38%)) using scanner 1, revealed observer 1 had mean (95% LOA) bias of 3.95 (-2.06 to 9.96) ml/min compared to 0.45 (-12.74 to 13.65) ml/min for observer 2. For scanner 2, bias was -2.42 (-12.18 to 7.34) ml/min for observer 1 and 0.54 (-17.00 to 18.09) ml/min for observer 2. Pulsatile flow measurements (n=30(62%)) using scanner 1, demonstrated that observer 1 had a bias of -0.63 (-3.82 to 2.57) ml/min compared to 0.82 (-5.88 to 7.52) ml/min for observer 2. Using scanner 2, bias was -2.49 (-6.18 to 1.21) ml/min for observer 1 and -2.60 (-7.95 to 2.75) ml/min for observer 2. Inter and intra-observer reliability was excellent for most pulsatile flow measurements.

We found excellent inter and intra-observer reliability for majority of pulsatile flow measurements and lower variance for observer 1. Researchers and clinicians can use 3D printed flow phantoms for training and assessing operator capabilities using different ultrasound scanners.