Spinal cord impingement by a displaced rib in scoliosis due to neurofibromatosis.

A 10-year-old boy was referred for evaluation of unusual backache with associated scoliosis. His back pain would be accentuated by lying on his stomach watching TV. When he rolled onto his right side, his right leg would get weak and he often felt a “shock” that would radiate to his toes. Neurofibromatosis type I had been diagnosed when he was 6 years old, when the skin lesions began to appear, but no spinal deformity was noted. When he presented to our scoliosis clinic, a right-sided prominent rib hump associated with a 76° thoracic scoliosis was seen. There was no evidence of any permanent neurologic impairment. Radiologic examination demonstrated a right thoracic curve from T4 to T12, measuring 76° by the Cobb method (Fig. 1). His curve had progressed from 55° to 76° over the preceding year. MRI showed no tumours intracranially or within the spinal canal. Threedimensional CT and CT-myelography revealed detachment and translocation of the cephalad end of the ninth rib on the convex side of the curve through an enlarged intervertebral foramen into the spinal canal, where it was severely compressing the spinal cord, even though the cord lay against the concave side of the spinal canal (Fig. 2, left). After thorough evaluation by both orthopedic and neurologic surgical teams, it was elected to perform a 2-stage procedure in an attempt to minimize postoperative complications such as paraparesis and paraplegia described in other reported cases. The first stage comprised rib excision with decompression of the protruding right ninth rib into the thoracic canal. We had planned to excise the whole posterior third of the rib. However, attempts to remove the rib from the canal resulted in a loss of signal by spinal cord monitoring. The rib was adherent to the cord. Because of this, the small intraspinal segment of the rib was left in situ, and a 5-cm segment starting at the nerve-root foramina was excised with the rib periosteum. The rib could then not exert a lever effect on the cord, and we hoped the remaining intraspinal portion would eventually be resorbed, similar to intraspinal bone fragments after burst fractures. After the procedure was completed, a wake-up test was performed, and the boy moved his feet well. Postoperatively, he was ambulatory the next day and exhibited no change in neurologic