Spatiotemporal Accuracy of Real-Time 3D Echocardiography in the Neonatal and Pediatric Setting – Validation Studies using Small Dynamic Test Objects

Abstract

The precision of real-time 3D-echocardiography (RT3DE) is not sufficiently validated for small, fast-moving structures such as the neonatal and pediatric heart. To assess the spatiotemporal accuracy of RT3DE in small, moving test objects. Small, calibrated test objects in the size of neonatal and pediatric heart chambers were made from polyurethane foam or metal wire mesh and moved in a water bath through a calibrated dynamic test system. Using matrix transducers (X7-2, ie33 and X4-1, Sonos 7500, Philips, Andover, USA), 2 D and live 3 D datasets under variation of the motion speed (0.033 - 0.133 m/s corresponding to 50 - 200 heart cycles/minute), the volume rate and transducer position were recorded and analyzed (QLab 7.0, Philips). 3 D datasets of the moving test objects showed relevant spatial distortion, which was obviously related to the sequential scanning technology of the matrix transducer. Different segments of a test object were not recorded simultaneously, but rather row-by-row, so that there was a time delay between the first and the last-recorded voxel of a single 3 D volume (mean±SD: 28.9 ± 7.82 m/s or 80 ± 7 % of the time duration of a 3 D volume). With increasing motion speed of the test object and reduced 3 D volume rate, the distortion artifacts increased significantly. 3 D acquisitions using matrix technology demonstrate relevant spatiotemporal inaccuracies. This may lead to misinterpretations during the evaluation of the synchronicity of valvular or ventricular motion and incorrect definition of volume estimations. In particular, at higher heart rates and higher rates of movement, these limitations have to be taken into account in clinical practice.