Antenatal Ultrasound and Magnetic Resonance Imaging in Localizing the Level of Lesion in Spina Bifida and Correlation With Postnatal Outcome

Abstract

The level of the lesion is the single most accurate predictor of the functional neurologic outcome in infants having spina bifida. The investigators retrospectively compared the results of the prenatal ultrasound examination and magnetic resonance imaging (MRI) with the postnatal surgical and MRI findings in a series of 12 antenatally and 5 postnatally diagnosed cases of spina bifida referred for surgical treatment in the years 2000 to 2002. Both the level of the lesions and the cranial head signs were compared with neurologic status after a mean follow up of 12 months. For the 12 parentally diagnosed cases, the median gestational age at the time of diagnostic ultrasound was 20 weeks and delivery was at 38 weeks. The 5 postnatally diagnosed cases also had an ultrasound examination at 20 weeks, but the spine lesions were not identified. Eight of the 12 women who had prenatal diagnosis and 2 of the 5 with postnatal diagnosis took folic acid periconceptionally. The MSAFP screening test was offered to 16 of the 17 study patients, but 8 of the 12 prenatally diagnosed women and all 5 with postnatal diagnosis declined. Three women with prenatal diagnosis accepted MSAFP testing and 2 were positive; the third had a fetus with a closed lumbar lesion. Four of the 12 (33%) prenatal cases and 3 of the 5 (60%) postnatal cases had high lesions with the lowermost part of the defect at or above L3, and 8 of the 12 prenatal (67%) and 2 of the 5 (40%) postnatal cases had low lesions with the uppermost part of the defect at or below L4. The antenatal findings correlated with the postnatal findings in 8 (67%) of the 12 antenatal cases; in 2 cases, ultrasound predicted a defect 2 vertebral bodies higher than it actually was, and in 2 cases, the defect was predicted to be 2 vertebral bodies lower. One or more head signs were seen sonographically in 10 of 12 antenatally diagnosed cases (83%). Seven infants had ventriculomegaly and Arnold-Chiari malformations confirmed in all cases by postnatal MRI. In contrast, although 8 of the 12 sonographically diagnosed cases also had antenatal MRI, the findings correlated with postnatal findings in only 4 cases and were inaccurate in 4 cases. There was a discrepancy between the antenatal ultrasound and MRI findings in 4 of 8 cases evaluated with both modalities; MRI was more accurate in 2 cases and ultrasound was more accurate in 2 cases. More than 90% of infants had motor deficits of varying severity. Neither the level of the spinal lesion nor the presence of ventriculomegaly correlated significantly with developmental milestones or with morbidity or motor function at follow up; 82% of all infants appeared to be developing normally regardless of the level of the defect. Low spinal lesions were associated with increased bladder morbidity compared with highly placed lesions (odds ratio, 10.0; 95% confidence interval, 1.05-95.01), but overall, the numbers of high and low lesions were too small to draw meaningful conclusions. In this study, antenatal ultrasonography was fairly accurate in diagnosing and characterizing open spine defects, and antenatal MRI did not add to the diagnostic success. The ultrasound findings, in fact, correlated better than did MRI with the surgical or postnatal MRI findings with respect to both the level of the lesion and head signs.