Center of rotation analysis for thoracic and lumbar 3-column osteotomies in patients with sagittal plane spinal deformity: insights in geometrical changes can improve understanding of correction mechanics

Abstract

OBJECTIVE Three-column osteotomy (3CO) is used for severe spinal deformities. Associated complications include sagittal translation (ST), which can lead to neurological symptoms. Mismatch between the surgical center of rotation (COR) and the concept of the ideal COR is a potential cause of ST. Matching surgical with conceptual COR is difficult with pedicle subtraction osteotomy (PSO) and vertebral column resection (VCR). This mismatch influences correction geometry, which can prevent maximum possible correction. The authors’ objective was to examine the sagittal correction geometry and surgical COR of thoracic and lumbar 3CO. METHODS In a retrospective study of patients with PSO or VCR for severe sagittal plane deformity, analysis of surgical COR was performed using pre- and postoperative CT scans in the PSO group and digital radiographs in the VCR group. Radiographic analysis included standard deformity measurements and regional kyphosis angle (RKA). All patients had 2-year follow-up, including neurological outcome. Preoperative CT scans were studied for rigid osteotomy sites versus mobile osteotomy sites. Additional radiographic analysis of surgical COR was based on established techniques superimposing pre- and postoperative images. Position of the COR was defined in a rectangular net layered onto the osteotomy vertebrae (OVs). RESULTS The study included 34 patients undergoing PSO and 35 undergoing VCR, with mean ages of 57 and 29 years and mean RKA corrections of 31° and 49°, respectively. In the PSO group, COR was mainly in the anterior column, and surgical and conceptual COR matched in 22 patients (65%). Smaller RKA correction (27° vs 32°, p = 0.09) was seen in patients with anterior eccentric COR. Patients with rigid osteotomy sites were more likely to have an anterior eccentric COR (41% vs 11%, p = 0.05). In the VCR group, 20 patients (57%) had single-level VCR and 15 (43%) had multilevel VCR. COR was mainly located in the anterior or middle column. Mismatch between surgical and conceptual COR occurred in 24 (69%) patients. Larger RKA correction (63° vs 45°, p = 0.03) was seen in patients with anterior column COR. Patients with any posterior COR had a smaller RKA correction compared to the rest of the patients (42° vs 61°, p = 0.007). CONCLUSIONS Matching the surgical with the conceptual COR is difficult and in this study failed in one- to two-thirds of all patients. In order to avoid ST during correction of severe deformities, temporary rods, tracking rods, or special instruments should be used for correction maneuvers.