Purpose: To compare transabdominal ultrasonography (US) with fetal magnetic resonance (MR) imaging in the prenatal evaluation of myelomeningocele lesion level.
Materials and methods: Prenatal US images, pre- and postnatal MR images, and postnatal spinal radiographs obtained in the first 100 fetuses who underwent intrauterine myelomeningocele repair were the basis for this study. Each image was used to assign a lesion level. The assigned levels were compared by means of the kappa statistic, as an index of agreement.
Results: All fetuses underwent prenatal US. Sixty-one fetuses underwent prenatal MR imaging. Fifty fetuses underwent both postnatal spinal radiography and postnatal MR imaging, and an additional 34 fetuses underwent one postnatal study but not the other. When findings on prenatal US images were compared with those on postnatal radiographs, the findings agreed within one spinal level in 79% (55 of 70, kappa = 0.60) of cases. When findings on prenatal MR images were compared with those on postnatal radiographs, the findings agreed in 82% (31 of 38, kappa = 0.63) of cases. Findings on postnatal MR images and those on postnatal spinal radiographs agreed within one spinal level in 100% (50 of 50, kappa = 1.0) of cases.
Conclusion: Findings at prenatal MR imaging and prenatal US are equally accurate for the assignment of a lesion level in a fetus with myelomeningocele. Given that findings with both modalities will lead to misdiagnosis of the spinal level by two or more segments in at least 20% of cases, care should be exercised when neurologic outcome is predicted on the basis of these studies alone.