Devoted to diagnostic and interventional spine imaging and therapeutics


MR and Ultrasound Assessment of the Level of Spinal Dysraphism in Fetuses Undergoing In Utero Myelomeningocele Repair 2005

General Spine

Deborah M Zarnow, MD, Non ASSR Member
Erin M Simon , MD, ASSR Member
Beverly G Coleman , MD, Non ASSR Member
Larissa T Bilaniuk , MD, Non ASSR Member
Enrico Danzer , MD, Non ASSR Member
Leslie N Sutton , MD, Non ASSR Member


Compare the level of spinal dysraphism diagnosed by prenatal ultrasound (US) with the level on postnatal MR in fetuses undergoing in utero myelomeningocele (MMC) repair to determine the accuracy of prenatal diagnosis, and the potential impact on the assessment of the utility of in utero MMC repair.

Methods & Materials

The level of dysraphism determined prospectively from the prenatal US report was compared with retrospective review of postnatal MR of the entire spine in 44 fetuses who underwent in utero MMC repair at our institution between 1998-2002 (24 female, 20 male; age range at US: 19 weeks, six days – 24 weeks, five days; MR: 32 weeks - 4 years, 11 months). On US, the level was assigned by counting inferiorly from the last rib (presumed to be arising from T12) or superiorly from the fourth sacral segment. On MR, the levels of posterior element dysplasia and frank deficiency were determined by counting the entire vertebral complement, from the craniocervical junction.


On US, 5 fetuses had thoracic lesions, 36 lumbar, and three sacral. On MR, three fetuses had a frank defect beginning in the thoracic spine, 35 lumbar, and 6 sacral. All spines had at least one level of posterior element dysplasia immediately cephalad of the defect on MR (range: 1-4 levels, mean: 1.68). Inclusion of the dysplastic, but not frankly deficient, levels changed the lesion distribution to 8 thoracic, 36 lumbar, 0 sacral. The US level was the same as the MR defect level in 22 cases (50%) and the same as the dysplastic level in 6, making it correct in 63.6% of cases. In an additional 6 cases, US was off by one level, resulting in clinically relevant accuracy in a total of 34/44 (77%). When comparing strictly to the frank defect, in 20 (45%), the US level was higher than the MR (by one level in 13 cases, 2 levels in 6 cases, 3 levels in one case). In 2 (5%), the US level was lower than the MR defect (by one level in one case, 2 levels in 1 case). On MR, 40 of 44 (90%) patients had a full complement of vertebrae, with 3 having only four lumbar-type bodies, and one having six. In three of these cases, there was discrepancy between the US level and MR level (MR lower in 2, higher in 1). A temporal/experiential pattern was also noted in the US diagnostic ability, as discrepancies were present 7 of the first 10 scans performed but only one of the last 10 scans performed.


In assessing the MMC level, consistency between prenatal US diagnosis and postnatal MR diagnosis was seen in the majority of cases in our series. Posterior element dysplasia above the level of the frank defect was universally identified with detailed postnatal MR evaluation. Consistency is required when interrogating prenatal US and MR for discrimination between dysplastic but present, and frankly deficient, posterior elements. Consistent definitions are required for US and MR to determine the true lesion level in these patients, as part of the assessment of efficacy of the in utero MMC repair relies upon the improvement in lower extremity function based on predictions from the level of the lesion. Correlation of MR findings, prenatal ultrasound results, and clinical status, including lower extremity function, is ongoing. While one level of discrepancy is currently clinically acceptable, with improved imaging technique we hope to improve diagnostic accuracy and counseling for prospective parents regarding neurological function.