Reliability and validity of three-dimensional fetal brain volumes

Objective. To determine the intraobserver and interobserver variability in calculating three-dimensional fetal brain volumes and to examine the relationship between these volumes and biparietal diameter and head circumference measurements and estimated gestational age. Methods. Eighty-five subjects between 16 and 40 completed weeks' gestation participated in the Institutional Review Board-approved study. Fetal head images were obtained axially and stored on a magnetic optical disk. The fetal brain volumes were calculated in triplicate by each of 2 observers using 8 to 10 coronal cuts. The coefficient of variation was determined for both physicians. Pearson correlations and linear regression were used to evaluate the relationship between three-dimensional head volume and standard biparietal diameter and head circumference measurements and estimated gestational age. Results. The coefficients of variation were low for both investigators, at 2.04% and 2.44%. The correlations between fetal brain volumes and biparietal diameter, head circumference, and estimated gestational age were all highly significant (P < .001). The linear regression of brain volumes with estimated gestational age was also highly significant (P < .001). Conclusions. Three-dimensional fetal brain volume measurements had excellent intraobserver and interobserver reliability. The volumes correlated very well with standard biparietal diameter and head circumference measurements. These volumes can also be used to determine estimated gestational age.