Purpose:

To evaluate how 3D parameters of the instrumented thoracic spine relate to the uninstrumented lumbar spine following thoracic-only fusion (TOF) for adolescent idiopathic scoliosis (AIS) and determine the factors predictive of lumbar correction.

Methods:

A multi-center retrospective review was conducted of AIS patients with Lenke 1-4 B or C curves undergoing fusion of their thoracic spine only with minimum 2-year follow-up and 3D spine reconstructions from biplanar radiography. Pre-operative to 2-year post-operative differences were evaluated. Pearson’s correlations were used to identify 3D coronal, sagittal and axial relationships at 2 years post-operative. Multivariate linear regression was used to identify variables most predictive of lumbar curve correction.

Results:

Eighty-four AIS patients met inclusion (54 B modifiers, 30 C modifiers). Average pre-operative thoracic and lumbar curves were 54 ± 9° and 41 ± 7° and corrected to 19 ± 7° and 21 ± 7°, respectively. 3D T5-T12 thoracic kyphosis increased from 6 ± 13° to 26 ± 8°. Thoracic and lumbar apical rotation decreased from – 14 ± 6° to -5 ± 6° and 13 ± 5° to 11 ± 6°, respectively. Thoracic curve correction correlated with lumbar curve correction (r = 0.67, p = 0.001) and decreased LIV tilt correlated with smaller residual lumbar curve magnitude (r = 0.63, p < 0.001). Increasing 3D kyphosis of the instrumented segment correlated with increased percentage lumbar curve correction (r = 0.29, p = 0.009). Multivariate linear regression showed LIV tilt and thoracic curve magnitude as the most predictive variables of post-operative residual lumbar curve magnitude, and percent correction of the thoracic curve and change in 3D instrumented kyphosis as most predictive of percent correction of the lumbar curve.

Conclusions:

Analysis of 3D data highlights several considerations for AIS patients undergoing TOF. Maximizing thoracic curve correction, leveling the LIV, and to a lesser extent, restoring kyphosis in the instrumented segment are the factors most likely to result in greater correction of the instrumented lumbar curve following thoracic-only fusions in Lenke 1-4 B and C curves.