Reliability of radiographic parameters in neuromuscular scoliosis

STUDY DESIGN. Retrospective review of radiographic data. OBJECTIVES. This study sought to define interoberserver and intraobserver variability to further delineate reliable means by which radiographs of patients with neuromuscular scoliosis can be examined. SUMMARY OF BACKGROUND DATA. Previous studies analyzed the use of Cobb angles in the measurement of idiopathic and congenital scoliosis, but no study until now describes a critical analysis of measurement in evaluating neuromuscular scoliosis. METHODS. Forty-eight patients with neuromuscular scoliosis radiographs were reviewed. These were evaluated for Cobb angle, end vertebrae selection, Ferguson angle, apex of the curve, C7 balance, pelvic obliquity, Risser sign, status of the triradiate cartilage, kyphosis Cobb angle, endplate selection for kyphosis, and kyphotic index. Interclass and intraclass variability was examined with statistical analysis. RESULTS. Cobb angle had an intraobserver variability was 5.7° and the interobserver variability was 14.8°. The intraobserver and interobserver variability for Ferguson angle was 6.8° and 20.6°, respectively. The kyphotic Cobb angle intraobserver variability was found to be 17.4°, and the interobserver variability was 24.01°. CONCLUSIONS. Neuromuscular scoliosis radiographs can be reliably analyzed with the use of Cobb angle. Other forms of analysis, such as Ferguson angle, are not as reliable. Pelvic obliquity should be measured from the horizontal, as other methods are not as reliable. Kyphosis is best evaluated with the use of the kyphotic Cobb angle. Finally, it is felt that a separate anteroposterior pelvis radiograph should be used to assess skeletal maturity, as scoliosis films often truncate the vital anatomy necessary to determine skeletal maturity.