Characterisation of torso segmental rotation changes after surgical correction of adolescent idiopathic scoliosis

INTRODUCTION A plethora of measurements are available to characterise scoliosis, but a mismatch remains between cosmesis, radiological parameters and patient functional outcomes. Better assessment of cosmesis, through quantitative techniques, may help address this mismatch. We propose a new method of quantifying external scoliotic deformity, by assessing rotation and symmetry in transverse slices across the length of the torso, using 3D photography. METHODS 3D surface scanning is conducted routinely at the Queensland Children’s Hospital spine deformity clinic, for patients with a diagnosis of Adolescent Idiopathic Scoliosis (AIS). From an ongoing database of pre- and post-operative scans a series of 6 patients with similar curve type (Lenke Type 1) were selected for analysis. Lenke Type 1 is defined as major thoracic structural curves, with any adjacent curves present being deemed compensatory. The mean age pre-operatively was 15.4 ± 1.6 years, and the mean major curve Cobb angle was 65.7 ± 13.3deg. 3D scans were captured (Artec Eva, Artec Group Inc., Luxembourg) at the preoperative surgical planning appointment, and again at either their immediate postoperative (6-8 weeks), or subsequent follow up appointments (6-12 months). Using the surface scan data, 3D virtual representations of the patient’s standing body shape were created (Figure 1). From these reconstructions, transverse cross-sectional profiles for the outer torso skin surface were created at 10% intervals between their pelvis and C7 prominence. A line of maximum symmetry was determined, that defined the angle of rotation of each profile (positive: right) relative to the sagittal plane. The range of profile rotations along the length of each individual patient’s spine (minimum to maximum) was compared pre- to post-operatively. Additionally the ratio of corresponding pixels about the line of maximum symmetry of each profile was assessed along the length of the spine, pre- and post-operatively. A healthy adult female with no spinal deformity, who underwent a 3D scan for the purposes of another research project was included as a reference. RESULTS AND DISCUSSION The selected patients saw a dramatic improvement in their major curve angle – the traditional common measure of surgical correction. The mean decrease was 42.2 ± 10.3 deg. A concurrent increase in standing height was also seen of between 0.2 – 5.6cm (mean 3.0cm). Torso profile rotations varied in magnitude between participants but the pattern was consistent and matched the curve type. Maximum rotation was seen at levels 4 and 5, approximately the T9-T11 rib region, corresponding with major curve apexes at T8-9 to T10-11. Decreases in the maximum rotation and the total range of rotations were seen post-operatively, particularly around the peak areas of profiles 4-6. In comparison, the healthy participant had a very consistent rotation of 3.4±0.5deg. All profiles included in the analysis had symmetry values between 90% and 100% (52 out of 53 were greater than 94%). Some proximal torso profiles that included the shoulders and neck were excluded because of clothing or postural issues. The healthy participant showed a mean symmetry value of 98.8% ± 0.60%. Increases in symmetry were seen in the study subjects after surgery. Profile 4 in particular showed the largest increase in symmetry (range 3.0% – 6.8%), with smaller changes found in the upper thoracic regions. CONCLUSIONS Torso segmental rotations and symmetry provide a metric to describe the externally visible trunk deformity associated with AIS. Main thoracic structural curves often involve a large degree of both anterior and posterior rib cage deformity that is not well characterised clinically. 3D photography has the potential to provide such a metric.