Quantitative analysis of the closure pattern of the neurocentral junction as related to preexistent rotation in the normal immature spine.

Abstract

The normal spine is not a symmetrical structure. In recent studies, we demonstrated the presence of an axial rotational pattern that is similar to what is seen in the most prevalent curve patterns in idiopathic scoliosis at different ages. This suggests that if the spine starts to decompensate into scoliosis, it follows this preexistent rotational pattern. In scoliosis, the neurocentral junctions (NCJs) close asymmetrically, which leads to a different pedicle morphology in the convexity and concavity of the curve. The present study aimed to establish at which age the NCJ closes in different regions of the spine, whether it closes asymmetrically in the nonscoliotic spine as well and whether the closure pattern is related to the earlier demonstrated preexistent rotation. To evaluate the closure pattern and surface area of the left and right NCJs throughout the normal immature spine in relation to the preexistent spinal rotation at different ages. Retrospective cohort study using a systematic, semiautomatic analysis. Computed tomography (CT) scans of the thorax and abdomen of 199 nonscoliotic children (0-16 years old) were systemically analyzed. CT scans had been obtained for several reasons unrelated to this study, for example, recurrent respiratory infections, malignant disease (not involving the spine), or work up before bone marrow transplantation. Scans were categorized according to the criteria of the Scoliosis Research Society into infantile (0-3 years old), juvenile (4-9 years old), and adolescent (10-16 years old) age cohorts. Closure, absolute surface area, and the angle between the longitudinal axis of the left and right NCJ and preexistent vertebral rotation at each spinal level. Transverse CT slices were systemically analyzed for closure and asymmetry of the absolute area of 4,992 NCJs from spinal levels T2-L5. The outcome measures were analyzed semiautomatically using custom-made software developed at our institution (ImageXplorer; Image Sciences Institute). Inter- and intraobserver reliabilities were calculated. For all subjects, the entire thoracic area was available. Complete scans down to L5 of the lumbar spine were available in 43 cases. Closure of the NCJs was first observed in the lumbar spine, then in the high thoracic spine, and finally in the mid- and low thoracic spine. Closure occurred asymmetrically, left-right predominance depended on the age. In the mid- and low thoracic spine, the surface areas of the right NCJs were larger at the infantile age, whereas at the juvenile age the areas of the left NCJs were larger. This corresponded to the spine's preexistent rotation. Rotation of the high thoracic vertebrae was to the left in all age cohorts. Rotation in the mid- and low thoracic spine was to the left in the infantile cohort but reversed to the right in the juveniles and even more so in the adolescents. The lumbar spine was rotated to the left at the infantile age and not significantly rotated at the juvenile and adolescent ages. Orientation of the NCJs in relation to the vertebraes' longitudinal axis was symmetrical, not dependent on age, and more transverse at the midthoracic levels than at other spinal levels. This study focuses on the asymmetry and the regional pattern of closure of the NCJs at different ages. It suggests that preexistent rotation of the spine is related to the asymmetrical closure of the NCJs. Whether the asymmetry is the cause of or is caused by the preexistent rotation cannot be derived from this study.