Assessment of corpus callosum biometric measurements at 18 to 32 weeks' gestation by 3-dimensional sonography.

OBJECTIVES The purposes of this study were to construct reference limits for corpus callosum dimensions measured on images reconstructed from 3-dimensional (3D) sonography and to evaluate the reproducibility of these measurements. METHODS Three-dimensional sonographic volumes were acquired transabdominally from an axial view of the head in 361 fetuses cross-sectionally studied at 18 to 32 weeks' gestation. Offline analysis of the fetal brain midsagittal plane was used to evaluate the length and area of the corpus callosum and corpus callosum-cavum septi pellucidi complex. The agreement between 2-dimensional (2D) and 3D measurements as well as the interobserver variability in 3D measurements were assessed by interclass correlation coefficients (ICCs). RESULTS Adequate visualization of the midsagittal plane was obtained in 98.1% of the fetuses. A clear distinction between the corpus callosum and cavum septi pellucidi was obtained in 35.7% of the fetuses, whereas in the remaining cases, the corpus callosum-cavum septi pellucidi complex was visualized as a single echogenic structure. The corpus callosum-cavum septi pellucidi complex length (r = 0.806; P < .0001), corpus callosum-cavum septi pellucidi complex area (r = 0.920; P < .0001), and corpus callosum area (r = 0.713; P < .0001) showed a significant linear growth with gestation. A good agreement was found between measurements from both 2D and 3D sonographic views (corpus callosum length ICC, 0.916) as well as between measurements obtained by different observers (corpus callosum-cavum septi pellucidi complex length ICC, 0.936; corpus callosum-cavum septi pellucidi complex area ICC, 0.931). CONCLUSIONS Measurements of the corpus callosum and cavum septi pellucidi can be obtained from the midsagittal plane of the fetal brain reconstructed from 3D volumes acquired transabdominally. The constructed nomograms may facilitate the diagnosis of corpus callosum abnormalities.