Interobserver reproducibility of transabdominal 3-dimensional sonography of the fetal brain

Objective. The purpose of this study was to assess the interobserver reproducibility of transabdominal 3-dimensional (3D) fetal neurosonography. Methods. This was a prospective observational study. We studied 23 consecutive singleton pregnancies between 18 and 23 weeks’ gestation. All cases had normal fetal neurosonographic examination findings, which were confirmed after birth. A 3D sonographic volume of the fetal head was acquired transabdominally by a single operator using an axial approach. Fetal brain anatomy was later analyzed offline by 2 different operators. Axial, sagittal, and coronal views of the fetal brain were obtained to perform a detailed evaluation of the fetal brain. Each operator defined the scanning planes obtained as adequate or inadequate. Results were evaluated with 2 × 2 tables and the Cohen κ coefficient to assess interobserver agreement. Results. Good-quality multiplanar images were obtained in 23 of 23 cases. The rate of adequate visualization was 100% for all of the axial planes, with κ values of 1.00. For sagittal and coronal planes, the rate of visualization ranged between 78% and 91%, with κ values ranging between 0.61 and 0.83. Conclusions. Transabdominal 3D sonography of the fetal brain at 18 to 23 weeks’ gestation has an acceptable degree of interobserver reproducibility. Key words: brain; interobserver agreement; neurosonography; 3-dimensional sonography.