Volumetry of fetal hearts using 3D real-time matrix echocardiography - in vitro validation experiments and 3D echocardiographic studies in fetuses

Abstract

Purpose: The aim of this study was to assess the feasibility, accuracy and reliability of 3D real-time echocardiography for fetal heart volumetry.

Materials and methods: Fifty unselected and consecutive fetuses, including 14 with cardiac malformations, were scanned prospectively using real-time 3D matrix technology and 2D echocardiography to determine ventricular volumes. Small phantoms as well as modified balloons (0.5 - 20 ml) were used to assess the validity of 2D and 3D distance, area and volume calculations and to study potential sources of error during data acquisition and analysis. The data was evaluated by two blinded observers.

Results: In vitro, real-time 3D and 2D underestimated the actual volumes by -5.49 % (3D) and -6.86 % (2D). The intraobserver and interobserver variability were excellent. In vivo, real-time 3D was superior to 2D with regard to intraobserver and interobserver variability (mean coefficient of variation 8.28 % (3D) versus 13.96 % (2D), and mean intraclass correlation coefficient 0.997 (3D) versus 0.885 (2D) for left ventricular volumes). Similar to in vitro, in vivo 2D volumes were calculated smaller than 3D volumes (mean difference -0.39 to -0.94 ml). The ventricular volumes and stroke volumes increased exponentially with gestation. Secondary to poor imaging windows in advanced gestation or inadequate delineation of endocardial borders in small hearts at less than 19 weeks, 3D data could not be analyzed sufficiently in 6 / 50 fetuses.

Conclusion: Real-time 3D using a matrix transducer is a feasible, reliable and valid method for volume determination in the fetus beyond 19 weeks of gestation. If compared to 2D, real-time 3D echocardiography provides improved accuracy of cardiac volumetry, decreases intraobserver and interobserver variability and is a promising tool for the accurate assessment of cardiac size and function.