Objective: To describe a prenatal diagnostic method for evaluating spina bifida by three-dimensional ultrasonography.
Methods: Two- and three-dimensional ultrasonography were used to determine the extent of vertebral defects among fetuses with spina bifida. Spinal levels were independently counted from the most caudal thoracic vertebra with a rib (e.g., 12th thoracic rib). A virtual cutting plane was manipulated through a volume-rendered spine to generate optimal multiplanar views for this blinded analysis. Prenatal diagnosis was compared with a postnatal analysis of bony spine defects derived from radiographic films or magnetic resonance imaging.
Results: Nine fetuses were examined by two-dimensional ultrasonography (21.8 +/- 3.4 menstrual weeks) and three-dimensional ultrasonography (22.8 +/- 4.4 menstrual weeks). For two-dimensional ultrasonography, the spinal level agreed to within 1 vertebral segment in 6 of 9 infants. In contrast, three-dimensional ultrasonography agreed to within 1 vertebral segment in 8 of 9 infants. Three fetuses had vertebral defect levels on two-dimensional ultrasonography that were 1.5 to 2 segments away from postnatal findings. The same fetuses had results that were within 1 vertebral segment on three-dimensional ultrasonography. Volume rendering showed splayed vertebral pedicles and disrupted vertebrae. An intact meningeal sac was easily rendered in 5 of 9 subjects.
Conclusions: Multiplanar views are generally more informative than rendered views for localizing bony defects of the fetal spine. The level of the defect on three-dimensional ultrasonography correlates well with those on two-dimensional ultrasonography and postnatal imaging studies. This approach may improve characterization of spina bifida by adding diagnostic information that is complementary to the initial assessment by two-dimensional ultrasonography.