Objective: To compare the consistency and reproducibility of fetal left ventricular stroke volume(LVSV) measured with two-dimensional (2D) ultrasonography with Doppler capability, 2D M-model, and four-dimensional (4D) with spatiotemporal image correlation (STIC) in normal fetuses.
Methods: The LVSV was recorded by two experienced operators using 2D Doppler, 2D M-model, and 4D STIC in a randomised order of 30 normal fetuses at 26 to 30 weeks of gestation. Intraclass correlation coefficient (ICC) and Bland-Altman plot were engaged to evaluate the repeatability and agreement.
Results: 1There were not significantly difference among the three methods(P>0.05); 2The intraclass correlation coefficient of 2D Doppler , 2D M-model, and 4D STIC for the LVSV were respectively 0.93 (95% confidence interval (CI) 0.86-0.95), 0.89 (95% CI 0.82-0.91),and 0.95 (95% CI 0.88-0.97); 3When comparing the LVSV measured with three methods, the intraclass coefficient were respectively 0.959(2D Doppler V.S. 2D M-model), 0.961(2D M-model V.S. 4D STIC) and 0.978(2D Doppler V.S. 4D STIC). 4Regarding the LVSV with 4D STIC had a smaller systematic bias and limits of agreement than those with 2D Doppler and 2D M-mode.
Conclusions: There is a good agreement between LVSV measured by 2D Doppler, 2D M-mode or 4D STIC. The 4D STIC represents better interobserver agreement than the others, and promises to become the method of choice.