Accuracy, agreement, and reliability of fetal cardiac measurements using 4-dimensional spatiotemporal image...

The purpose of this study was to evaluate the accuracy, agreement, and reliability of 4-dimensional sonography using spatiotemporal image correlation (STIC) in fetal cardiac measurements during the second and third trimesters.

Four-dimensional cardiac STIC volumes were acquired from 150 low-risk singleton pregnancies at gestational ages of 18 to 38 weeks. A total of 11 dimensions of the fetal heart and great vessels were analyzed by two different observers. Accuracy was assessed by comparing the measurements acquired by 4-dimensional STIC with those determined by 2-dimensional echocardiography by calculating the mean error and mean percent error and depicted by Bland-Altman plots. Absolute agreement and reliability of STIC-based measurements were evaluated on the basis of an intraclass correlation coefficient (ICC).

For STIC-based measurements of cardiac dimensions, good accuracy was achieved by the two experienced observers (mean percent error, -3.10% to 0.87% for observer A; -2.49% to -0.33% for observer B), and absolute agreement was almost perfect (ICC, 95% confidence interval, 0.916-0.993 for observer A; 0.957-0.990 for observer B). Interobserver ICCs among the two observers were between 0.909 and 0.990 (95% confidence interval), whereas intraobserver ICCs for observer A were between 0.894 and 0.989, and those for observer B were between 0.893 and 0.985, showing no significant differences in the reliability of SITC with regard to observers.

Four-dimensional sonography with STIC is a feasible and accurate method for fetal cardiac dimension measurements in the second and third trimesters. It is in good agreement with 2-dimensional echocardiography and can be performed by different observers.