Intra‐ and interobserver repeatability of fetal cardiac examination using four‐dimensional spatiotemporal image correlation in each trimester of pregnancy

Abstract

To assess the intra- and interobserver repeatability of the evaluation of fetal cardiac structures and measurements using spatiotemporal image correlation (STIC) technology in each trimester of pregnancy. Four-dimensional (4D)-STIC volumes from 150 low-risk pregnancies were acquired at first-, second- or third-trimester scan for later analysis by two different reviewers. A total of 19 items, including the evaluation of 14 structures as well as five measurements of the fetal heart, were evaluated. The reliability of qualitative variables was evaluated using Cohen's kappa and absolute agreement analysis while that of quantitative parameters was assessed using the intraclass correlation coefficient (ICC). Forty-five, 47 and 47 STIC volumes were included in the final analysis from the first, second and third trimesters, respectively. For the evaluation of cardiac structures, good or excellent intra- and interobserver agreement (kappa > 0.6) was obtained in 12/14 and 9/14, respectively, while absolute agreement was > 90% for most structures evaluated. Regarding the quantitative cardiac measurements, ICC values were above 0.90 for all but cardiac axis (intraobserver ICC, 0.81 and interobserver ICC, 0.61). There were no significant differences in the repeatability values observed for qualitative or quantitative parameters among the trimesters of pregnancy. Cardiac examination from 4D-STIC volumes showed a high repeatability between and within observers in each trimester of pregnancy.