ASE: 3D Echo Tops 2D and M-Mode for Kids' Hearts

June 21, 2007

SEATTLE -- When assessing the left ventricle in children, real-time three-dimensional echocardiography offers the best combination of accuracy and efficiency, according to a small trial reported here.

SEATTLE, June 21 -- When assessing the left ventricle in children, real-time three-dimensional echocardiography offers the best combination of accuracy and efficiency, according to a small trial reported here.

Measurement of left-ventricular size, function, and mass is the most common and essential application of echocardiograph in pediatric clinical practice and research, said Xiluzhang Lu, M.D., of Baylor in Houston, and colleagues, at the American Society of Echocardiography meeting.

However, little data have existed on the relative accuracy, reproducibility, and efficiency of M-mode, 2D, and 3D echo techniques for acquiring left ventricular measurements, Dr. Lu added.

The Baylor investigators prospectively evaluated 19 consecutive healthy children (mean age 10.6 years) by means of M-mode, 2D, and RT3D echo, using the same type of machine for each evaluation, the Sonos 7500 by Philips Medical Systems. M-mode and 2D measurements were made in accordance with ASE recommendations. Real-time 3D measurements were acquired both manually and with a semi-automated algorithm. In addition, all of the children were evaluated by magnetic resonance imaging.

They found that M-mode echo is faster than 2D or real-time 3D echo but is the least accurate and least reproducible of the three modalities. Real-time 3D outperforms 2D echo for image accuracy and reproducibility, and use of an automated algorithm makes it as fast as 2D, giving the nod for best all-around imaging technique in children to real-time 3D.

"Three-D, using either automated or manual algorithm analysis, is superior to 2D for measurements of left-ventricular volumes, ejection fraction, and mass," said Dr. Lu. "The automated 3D echo algorithm is as efficient as 2D, and therefore, is the method of choice for quantification of left-ventricular indices."

The correlation and agreement between echo and MRI measurements were good or excellent, said Dr. Lu. The correlation and agreement of real-time 3D were superior to those of M-mode and 2D echo, and the manual and semi-automated analysis methods for real-time 3D produced comparable results.

For reproducibility, the intra- and inter-observer variability were similar for M-mode and 2D echo and superior with RT3D:

  • M-mode (4.3-4.8%, 7.0-8.7%)
  • 2D echo (3.3-4.5%, 5.5-7.3%)
  • Real-time 3D (0.4-2.3%, 0.2-4.8%)

M-mode required the least time for data acquisition and analysis (twp minutes). RT3D took less time for data acquisition compared with 2D but required more time for data analysis, resulting in total times of 20.5 minutes for and 26.2 minutes for four-plane and eight-plane manual algorithms, respectively, compared with 12.3 minutes for 2D. With the semi-automated algorithm, RT3D total time for data acquisition and analysis decreased to 11.1 minutes.