Abstract
BACKGROUND: Standing whole spinal radiographs are used to evaluate spinal alignment in adult spinal deformity (ASD), yet some studies have reported that pelvic incidence, pelvic tilt, and thoracic kyphosis (TK) intra- and inter-observer reliability is low. This study aimed to evaluate the accuracy of spinopelvic parameters through comparing standing whole spinal radiographs and upright CT images. METHODS: We enrolled 26 patients with ASD. All standing whole spinal posterior/anterior and lateral radiographs and upright whole spinal CT had been obtained in a natural standing position. Two examiners independently measured 13 radiographic parameters. Interclass correlation coefficients (ICCs) were used to analyze measurement intra- and inter-observer reliability. Paired t- and Pearson's correlation tests were used to analyze validity of the standing whole spinal radiographs. RESULTS: ICCs of upright CT were excellent in both intra- and inter-observer reliability. However, intra-observer ICCs for TK2-12, TK1-5, TK2-5, and TK5-12 on standing lateral radiographs were relatively low, as were inter-observer ICCs for TK2-12, TK1-5, TK2-5, and TK5-12. Concerning TK values, the difference between the radiographs and CT in TK1-12 and TK2-12 were 4.4 ± 3.1 and 6.6 ± 4.6, respectively, and TK values from T2 showed greater measurement error (p < 0.05). CONCLUSIONS: Upright CT showed excellent intra- and inter-observer reliability in the measurement of spinopelvic parameters. Measurement of TK with T2 on standing whole spinal radiographs resulted in a greater measurement error of up to 6.6°. Surgeons need to consider this when planning surgery and measuring postoperative TK changes in patients with ASD.