Automated Measurement of Stroke Volumes by Real-Time Three-Dimensional Doppler Echocardiography: Coming of Age?

Color Flow Ultrasound Spatial Sampling Beam Density for Partial Volume-Corrected Three-Dimensional Volume Flow (3DVF): Theory, Simulation, and Experiment

OBJECTIVE

The purpose of this study was to quantify the accuracy of partial volume-corrected three-dimensional volume flow (3DVF) measurements as a function of spatial sampling beam density using carefully-designed parametric analyses in order to inform the target applications of 3DVF.

METHODS

Experimental investigations employed a mechanically-swept curvilinear ultrasound array to acquire 3D color flow (6.3 MHz) images in flow phantoms consisting of four lumen diameters (6.35, 4.88, 3.18 and 1.65 mm) with volume flow rates of 440, 260, 110 and 30 mL/min, respectively. Partial volume-corrected three-dimensional volume flow (3DVF) measurements, based on the Gaussian surface integration principle, were computed at five regions of interest positioned between depths of 2 and 6 cm in 1 cm increments. At each depth, the color flow beam point spread function (PSF) was also determined, using in-phase/quadrature data, such that 3DVF bias could then be related to spatial sampling beam density. Corresponding simulations were performed for a laminar parabolic flow profile that was sampled using the experimentally-measured PSFs. Volume flow was computed for all combinations of lumen diameters and the PSFs at each depth.

RESULTS

Accurate 3DVF measurements, i.e., bias less than ±20%, were achieved for spatial sampling beam densities where at least 6 elevational color flow beams could be positioned across the lumen. In these cases, greater than 8 lateral color flow beams were present. PSF measurements showed an average lateral-to-elevational beam width asymmetry of 1:2. Volume flow measurement bias increased as the color flow beam spatial sampling density within the lumen decreased.

CONCLUSION

Applications of 3DVF, particularly those in the clinical domain, should focus on areas where a spatial sampling density of 6 × 6 (lateral x elevational) beams can be realized in order to minimize measurement bias. Matrix-based ultrasound arrays that possess symmetric PSFs may be advantageous to achieve adequate beam densities in smaller vessels.

Evaluation of a new 3-dimensional color Doppler flow method to quantify flow across the mitral valve and in the left ventricular outflow tract: an in vitro study

OBJECTIVES

The aim of this study was to assess the accuracy, feasibility, and reproducibility of determining stroke volume from a novel 3-dimensional (3D) color Doppler flow quantification method for mitral valve (MV) inflow and left ventricular outflow tract (LVOT) outflow at different stroke volumes when compared with the actual flow rate in a pumped porcine cardiac model.

METHODS

Thirteen freshly harvested pig hearts were studied in a water tank. We inserted a latex balloon into each left ventricle from the MV annulus to the LVOT, which were passively pumped at different stroke volumes (30-80 mL) using a calibrated piston pump at increments of 10 mL. Four-dimensional flow volumes were obtained without electrocardiographic gating. The digital imaging data were analyzed offline using prototype software. Two hemispheric flow-sampling planes for color Doppler velocity measurements were placed at the MV annulus and LVOT. The software computed the flow volumes at the MV annulus and LVOT within the user-defined volume and cardiac cycle.

RESULTS

This novel 3D Doppler flow quantification method detected incremental increases in MV inflow and LVOT outflow in close agreement with pumped stroke volumes (MV inflow, r = 0.96; LVOT outflow, r = 0.96; P < .01). Bland-Altman analysis demonstrated overestimation of both (MV inflow, 5.42 mL; LVOT outflow, 4.46 mL) with 95% of points within 95% limits of agreement. Interobserver variability values showed good agreement for all stroke volumes at both the MV annulus and LVOT.

CONCLUSIONS

This study has shown that the 3D color Doppler flow quantification method we used is able to compute stroke volumes accurately at the MV annulus and LVOT in the same cardiac cycle without electrocardiographic gating. This method may be valuable for assessment of cardiac output in clinical studies.