Purpose: To retrospectively determine the relationship between posterior fossa volume (PFV) and estimated gestational age (EGA) and/or femur length (FL) during pregnancy for the purpose of developing a normal growth curve.
Materials and methods: Advance institutional review board approval was obtained for this HIPAA-compliant study, and the need for parent informed consent was waived. A cross-sectional retrospective study was performed to measure PFV on in vivo magnetic resonance (MR) images obtained in 76 fetuses of 18-36 weeks gestation who had a morphologically normal CNS. Because this was a retrospective series, MR imaging techniques varied slightly, but all fetuses underwent imaging at contiguous 3-5-mm intervals in at least two orthogonal planes, with repetition time msec/echo time msec, 5-12/62-95; number of signals acquired, one; flip angle, 150 degrees -180 degrees; and matrix, 128-192 x 256. Posterior fossa areas were manually traced on half-Fourier rapid acquisition with relaxation enhancement in utero fetal MR images by one observer. PFVs were then calculated by manually summing areas from the contiguous sections and multiplying the total area by the section thickness. An average PFV (APFV) across orthogonal planes was calculated for each fetus, and the relationship between APFV and EGA was mathematically modeled. Coronal, transverse, and sagittal views were compared with correlations and Bland-Altman plots. Two additional observers repeated the measurements for a small subset of fetuses (n = 5). Paired t test analyses were also performed to determine significant differences between sagittal, transverse, and coronal measurements, as well as to determine preliminary intraobserver and interobserver variability of measurements in a subset of cases.
Results: The relationship between APFV (in cubic centimeters) and EGA (in weeks) was well described by a single exponential function [APFV = 0.689 exp(EGA/9.10)]. APFV doubling time was 6.31 weeks. Root-mean-square variation of values around the model line was 1.63 cm(3). There was no statistically significant intra- or interobserver variation (P > .16 for all fetuses) at preliminary analysis. No correlation between APFV and FL could be found.
Conclusion: The normal fetal PFV growth curve generated in this study may have potential as a model for clinical application.
Copyright RSNA, 2006.