Curve length, curve form, and location of lower-end vertebra as a means of identifying the type of scoliosis.

PURPOSE To determine if the curve length, curve form, and location of the lower-end vertebra can identify the type of scoliosis. METHODS Standing posteroanterior and lateral radiographs of 78 women and 27 men with scoliosis aged 8 to 32 years were retrospectively analysed. Parameters measured were (1) the curve length (the number of vertebrae in the main curve), (2) the curve form (C-form, inverted C-form, or S-form), (3) the curve apex (the vertebral body at the apex of each curve), (4) the site of the scoliosis (thoracic, thoracolumbar, or lumbar), and (5) the location of the lower-end vertebra. RESULTS Only 3 (4%) of 77 patients with idiopathic scoliosis had a curve length of >8 vertebrae, compared with 19 (90%) of 21 patients with neuromuscular/neuropathic scoliosis (p<0.001, Fisher's exact test). 14 (18%) of the patients with idiopathic scoliosis had an S-form curve, compared with none in those with neuromuscular/neuropathic or congenital/osteogenic scoliosis (p=0.035, Fisher's exact test). No patient with idiopathic scoliosis had the lower- end vertebra located at L5, compared with 8 (38%) patients with neuromuscular scoliosis (p<0.001, Fisher's exact test). The criteria for neuromuscular/neuropathic scoliosis (a curve length of >8 vertebrae, a C-form or inverted C-form curve, and a lower-end vertebra located at L4 or L5) had 81% specificity, 76% sensitivity, 100% positive and 93% negative predictive values. CONCLUSION These criteria may help identify patients with neuromuscular/neuropathic scoliosis for further investigation by magnetic resonance imaging or computed tomography before they undergo corrective surgery.