[Similar coronal curvature may not represent the same 3-dimensional deformity in adolescent idiopathic scoliosis: a matched-pair study using EOS imaging system].

To analyze the preoperative axial plane and the surgical outcomes of the Lenke type 1A patients with adolescent idiopathic scoliosis (AIS) whose coronal curve type was matched but thoracic kyphosis (TK) was different. This study retrospectively reviewed a series of Lenke type 1A female AIS patients who underwent corrective surgery in the Department of Spine Surgery of Nanjing Drum Tower Hospital from May to August 2017. After matched with the Cobb angle of the main thoracic curve, the apical vertebral, the vertebra number included in the curve, the lumber modifier in the Lenke classification and Risser sign, 12 pairs of AIS patients, whose coronal curve was matched but thoracic kyphosis was different, were included in this study. The patients were divided into normal TK group and thoracic hypokyphosis group. EOS whole-body images were taken preoperatively and reconstructed by three-dimensional reconstruction. The whole spine anteroposterior X-ray was taken at 3 weeks after surgery. The radiographic parameters were measured on the preoperative and postoperative two-dimensional X-ray images: coronal Cobb angle, TK, lumbar lordosis (LL), pelvic incidence (PI) and pelvic tilt (PT). The vertebra rotation was obtained on the EOS three-dimensional reconstructed image, and the average vertebral rotation of the major thoracic curve (MTR), the average vertebral rotation of the proximal thoracic curve (PTR) and the average vertebral rotation of the lumbar curve (LR) were calculated. The paired sample test was used to compare the difference of preoperative and postoperative radiographic parameters between the groups. A total of 24 patients (12 pairs) were included in this study with an average age of (13.7±2.9) years. The preoperative Cobb angle was similar in the two groups (53.8°±10.2° vs 51.0°±11.1°, =0.27, =0.81). The average preoperative TK of the normal TK groups was 28.2°±6.1°, while that of the thoracic hypokyphosis group was 11.2°±5.6°(=7.68, <0.01). The MTR in the normal TK group was significantly smaller than that in the thoracic hypokyphosis group (10.2° vs 12.7°, =-3.74, <0.01), and there was a significant correlation between TK and MTR (=0.30, =0.03). As for the lumbar curve rotation, the LR of the normal TK group was significantly smaller than that in the thoracic hypokyphosis group (=-2.65, =0.002), but the absolute value of the two groups was similar (=-0.33, =0.31). The lumbar Cobb angle correction rate was significantly greater in patients with thoracic and lumbar curve rotating in the same direction than that in the opposite direction (81.1% vs 61.9%, =4.24, =0.005). It indicated that when the coronal deformity is matched, the MTR of the patients with thoracic hypokyphosis is significantly larger than that in the patients with normal thoracic kyphosis. The direction of the thoracic and lumbar curve rotation is required to be well concerned in the preoperative surgical planning.