Joseph H Carreau1, Tracey Bastrom2, Maty Petcharaporn2, Caitlin Schulte2, Michelle Marks2, Tamás Illés3, Szabolcs Somoskeöy4, Peter O Newton5. 1. Department of Orthopaedics, University of California, San Diego, 9500 Gilman Drive, San Diego, CA 92093, USA. 2. Rady Children's Hospital and Health Center, San Diego, 3020 Children's Way, San Diego, CA 92123, USA. 3. Orthopedic and Trauma department in CHU - Brugmann, Université Libre de Bruxelles, Brussels, Belgium. 4. Department of Orthopedic Surgery Institute of Musculoskeletal Surgery, University of Pécs, 1 Akác utca, Pécs H-7632, Hungary. 5. Rady Children's Hospital and Health Center, San Diego, 3020 Children's Way, San Diego, CA 92123, USA. Electronic address: pnewton@rchsd.org.
Abstract
STUDY DESIGN: Reproducibility study of SterEOS 3-dimensional (3D) software in large, idiopathic scoliosis (IS) spinal curves. OBJECTIVE: To determine the accuracy and reproducibility of various 3D, software-generated radiographic measurements acquired from a 2-dimensional (2D) imaging system. SUMMARY OF BACKGROUND DATA: SterEOS software allows a user to reconstruct a 3D spinal model from an upright, biplanar, low-dose, X-ray system. The validity and internal consistency of this system have not been tested in large IS curves. METHODS: EOS images from 30 IS patients with curves greater than 50° were collected for analysis. Three observers blinded to the study protocol conducted repeated, randomized, manual 2D measurements, and 3D software generated measurements from biplanar images acquired from an EOS Imaging system. Three-dimensional measurements were repeated using both the Full 3D and Fast 3D guided processes. A total of 180 (120 3D and 60 2D) sets of measurements were obtained of coronal (Cobb angle) and sagittal (T1-T12 and T4-T12 kyphosis; L1-S1 and L1-L5; and pelvic tilt, pelvic incidence, and sacral slope) parameters. Intra-class correlation coefficients were compared, as were the calculated differences in values generated by SterEOS 3D software and manual 2D measurements. The 95% confidence intervals of the mean differences in measures were calculated as an estimate of reproducibility. RESULTS: Average intra-class correlation coefficients were excellent: 0.97, 0.97, and 0.93 for Full 3D, Fast 3D, and 2D measures, respectively (p = .11). Measurement errors for some sagittal measures were significantly lower with the 3D techniques. Both the Full 3D and Fast 3D techniques provided consistent measurements of axial plane vertebral rotation. CONCLUSIONS: SterEOS 3D reconstruction spine software creates reproducible measurements in all 3 planes of deformity in curves greater than 50°. Advancements in 3D scoliosis imaging are expected to improve our understanding and treatment of idiopathic scoliosis.
STUDY DESIGN: Reproducibility study of SterEOS 3-dimensional (3D) software in large, idiopathic scoliosis (IS) spinal curves. OBJECTIVE: To determine the accuracy and reproducibility of various 3D, software-generated radiographic measurements acquired from a 2-dimensional (2D) imaging system. SUMMARY OF BACKGROUND DATA: SterEOS software allows a user to reconstruct a 3D spinal model from an upright, biplanar, low-dose, X-ray system. The validity and internal consistency of this system have not been tested in large IS curves. METHODS: EOS images from 30 IS patients with curves greater than 50° were collected for analysis. Three observers blinded to the study protocol conducted repeated, randomized, manual 2D measurements, and 3D software generated measurements from biplanar images acquired from an EOS Imaging system. Three-dimensional measurements were repeated using both the Full 3D and Fast 3D guided processes. A total of 180 (120 3D and 60 2D) sets of measurements were obtained of coronal (Cobb angle) and sagittal (T1-T12 and T4-T12 kyphosis; L1-S1 and L1-L5; and pelvic tilt, pelvic incidence, and sacral slope) parameters. Intra-class correlation coefficients were compared, as were the calculated differences in values generated by SterEOS 3D software and manual 2D measurements. The 95% confidence intervals of the mean differences in measures were calculated as an estimate of reproducibility. RESULTS: Average intra-class correlation coefficients were excellent: 0.97, 0.97, and 0.93 for Full 3D, Fast 3D, and 2D measures, respectively (p = .11). Measurement errors for some sagittal measures were significantly lower with the 3D techniques. Both the Full 3D and Fast 3D techniques provided consistent measurements of axial plane vertebral rotation. CONCLUSIONS: SterEOS 3D reconstruction spine software creates reproducible measurements in all 3 planes of deformity in curves greater than 50°. Advancements in 3D scoliosis imaging are expected to improve our understanding and treatment of idiopathic scoliosis.