Generation of a Patient-Specific Model of Normal Sagittal Alignment of the Spine

Abstract Study Design Mathematical modeling of normal sagittal spinal alignment. Objective To create a patient specific 3-dimensional (3D) model of normal adolescent spinal shape and alignment. Summary of Background Data Recreating normal sagittal balance is a key goal in spinal deformity surgery. Because of the variation in normal sagittal alignment based on inherent pelvic parameters, it is difficult to know what is normal for a given patient who presents with spinal deformity. Methods Simultaneous biplanar 2-dimensional digital radiographs were taken for pediatric patients with no known spinal disease using the EOS system. Three-dimensional reconstructions were produced using sterEOS and imported into custom MATLAB software. The researchers defined relationships to approximate orientations and positions of the vertebral bodies from patients' pelvic incidence (PI). The predicted spinal contour was then calculated to optimize congruence to patients' sagittal T1–sacrum offset, sagittal curve inflection point location, and predicted vertebral body orientations and positions. Results A total of 75 patients (26 male and 49 female) were included, mean age 14.5 ± 2.6 years. Baseline measurements were PI 46.7° ± 10.2°, sacral lope 40.2° ± 8.9°, T1–T12 kyphosis 39.8° ± 8.8°, and L1–L5 lordosis −37.1° ± 11.2°. Average difference in vertebral position in the anteroposterior direction between actual spines and their predicted models was 1.2 ± 1.2 mm and varied from an absolute minimum of 0.2 mm (T3) to an absolute maximum of 3.7 mm (L2). Conclusions This model uses an adolescent patient's PI to predict the normal sagittal alignment that best matches that patient's native sagittal curve. The model was validated on patients with no spinal deformity; average difference between actual sagittal positions of each vertebra and those predicted by the model was less than 5 mm at each vertebral level. This model may be useful in adolescent scoliotic patients with altered sagittal alignment to determine the magnitude of 3D deformity (compared with predicted normal values) and the completeness of 3D correction.