Use of magnetic resonance imaging to detect occult spinal dysraphism in infants.

OBJECTIVE Cutaneous stigmata or congenital anomalies often prompt screening for occult spinal dysraphism (OSD) in asymptomatic infants. While a number of studies have examined the results of ultrasonography (US) screening, less is known about the findings when MRI is used as the primary imaging modality. The object of this study was to assess the results of MRI screening for OSD in infants. METHODS The authors undertook a retrospective review of all infants who had undergone MRI of the lumbar spine to screen for OSD over a 6-year period (September 2006-September 2012). All images had been obtained on modern MRI scanners using sequences optimized to detect OSD, which was defined as any fibrolipoma of the filum terminale (FFT), a conus medullaris ending at or below the L2-3 disc space, as well as more complex lesions such as lipomyelomeningocele (LMM). RESULTS Five hundred twenty-two patients with a mean age of 6.2 months at imaging were included in the study. Indications for imaging included isolated dimple in 235 patients (45%), asymmetrically deviated gluteal cleft in 43 (8%), symmetrically deviated (Y-shaped) gluteal cleft in 38 (7%), hemangioma in 28 (5%), other isolated cutaneous stigmata (subcutaneous lipoma, vestigial tail, hairy patch, and dysplastic skin) in 31 (6%), several of the above stigmata in 97 (18%), and congenital anomalies in 50 (10%). Twenty-three percent (122 patients) of the study population had OSD. Lesions in 19% of these 122 patients were complex OSD consisting of LMM, dermal sinus tract extending to the thecal sac, and lipomeningocele. The majority of OSD lesions (99 patients [81%]) were filar abnormalities, a group including FFT and low-lying conus. The rate of OSD ranged from 12% for patients with asymmetrically deviated gluteal crease to 55% for those with other isolated cutaneous stigmata. Isolated midline dimple was the most common indication for imaging. Among this group, 20% (46 of 235) had OSD. There was no difference in the rate of OSD based on dimple location. Those with OSD had a mean dimple position of 15 mm (SD 11.8) above the coccyx. Those without OSD had a mean dimple position of 12.2 mm (SD 19) above the coccyx (p = 0.25). CONCLUSIONS The prevalence of OSD identified with modern high-resolution MRI screening is significantly higher than that reported with US screening, particularly in patients with dimples. The majority of OSD lesions identified are FFT and low conus. The clinical significance of such lesions remains unclear.