Spinal sonography and magnetic resonance imaging in patients with repaired myelomeningocele: comparison of modalities.

The goals of this study were to evaluate the feasibility of using ultrasonography of the spine in the follow-up evaluation of patients with repaired myelomeningocele at birth and to compare sonography with the accepted modality of magnetic resonance imaging. Over a period of 4 years we performed 165 sonographic studies in 101 patients; 107 sonographic studies had MR imaging results for comparison. We collected our data prospectively. The quality of the sonograms was good in 110 of 129 studies, acceptable in 17 of 129, and poor in two of 129. The sonographic examinations failed in 33 of 165 studies (20%). Concordant information was obtained between ultrasonography and magnetic resonance imaging in the following percentage of studies: level of the distal end of the cord in 82%, position of the cord in the canal in 59%, presence of hydromyelia in 63%, cord duplication in 96%, adhesions in 16%, intradural mass in 37%, cord measurements in 85%, and dural sac measurements in 83%. At the lumbosacral level, we saw no cord pulsation in 57% of the studies in patients with cord adhesions and in 20% of those without adhesions. At the lower thoracic level, we saw no pulsation in 35% of the studies in patients with cord adhesions and in 7% of those without adhesions. Postoperative studies of cord release surgery in eight patients showed varied findings. We conclude that in those patients who have a spinal defect or interlaminar space allowing proper visualization of the lumbosacral spinal canal, ultrasound can provide fairly similar information to that obtained with magnetic resonance imaging of that area with no need for sedation and at a reduced cost. Ultrasonography seems more sensitive than magnetic resonance imaging in the detection of cord adhesions, which is particularly relevant in the diagnosis of tethering.