Abstract / Summary
Abstract Purpose Current spinal deformity diagnosis relies on whole-spine radiographs, while Weight-bearing Radiology 3D (WR3D) offers accurate, low-dose three-dimensional reconstruction in the standing position. This study aims to evaluate the potential of digitally reconstructed radiographs (DRRs) generated from WR3D images as a substitute for standard whole-spine radiographs. Methods In this retrospective comparative study of 67 patients who underwent both WR3D examination and whole-spine radiographs, DRRs were generated using segmentation and projection algorithms. Two observers independently measured coronal and sagittal radiographic parameters of the spine and pelvis on both imaging modalities, with intraclass correlation coefficients (ICCs) calculated. Paired samples t -tests and Bland-Altman analyses were performed, with mean absolute errors (MAEs) and Pearson correlation coefficients calculated. Results DRRs exhibited superior interobserver reliability compared to radiographs, particularly for pelvic incidence (PI, ICC = 0.95 vs. 0.70), pelvic tilt (PT, ICC = 0.98 vs. 0.83), sacral slope (SS, ICC = 0.94 vs. 0.79), and upper thoracic parameters. High agreement between the two imaging modalities was observed in measurements of all coronal parameters, cervical lordosis (C2-C7), thoracolumbar kyphosis (T10-L2), lumbar lordosis (L4-S1, L1-S1), PI, and PT ( p > 0.05), while differences were observed for other sagittal parameters ( p < 0.05). Conclusions DRR images based on WR3D show high agreement with radiographs for measuring most structural parameters. However, certain differences exist for global balance parameters and sagittal curvature parameters.