Objective: The purpose of this study was to evaluate the accuracy of 3-dimensional (3D) sonography in assessing fetal anatomy and to determine the intraobserver reproducibility and the effect of examiner experience.
Methods: Three-dimensional volumes of the head, face, thorax, and abdomen were obtained for 40 fetuses. The volume data sets obtained were explored offline with multiplanar navigation and tomographic ultrasound imaging on a personal computer. Each case was examined twice by the same observer at least 3 months apart. The percentage for identification of fetal anatomic structures, 2-dimensional (2D) and 3D measurements, and the time spent on 2D and 3D examinations were calculated and compared.
Results: Ninety-two percent of fetal anatomic structures were identified with multiplanar navigation and tomographic ultrasound imaging. The genitals, the entry of the vena cava, and the ears were visualized in less than 70% of cases. Tomographic ultrasound imaging allowed viewing of 14 structures not seen in the multiplanar study. Intraobserver agreement for anatomic examinations was good (kappa = 0.78). Intraobserver agreement for fetal measurements showed differences between both examinations that approached 0 and intraclass correlation indices close to 0.9. The mean 2D scanning time +/- SD was 10.11 +/- 2.5 minutes, and the acquisition time for the 3D volumes was 1.54 +/- 0.35 minutes. The total time for the second 3D study was 7.23 +/- 1.17 minutes, significantly shorter than the 9.96 +/- 1.53 minutes spent on the first study (P < .001).
Conclusions: Three-dimensional volumes are highly effective for complete fetal anatomic surveys. They show excellent intraobserver reproducibility and take less time to study as the examiner's experience increases.