• Laure Boyer, Jesse Shen, Stefan Parent, Samuel Kadoury, and Carl-Eric Aubin.
• Polytechnique Montreal, P.O. Box 6079, Downtown Station, Montreal, Quebec H3C 3A7, Canada; Research Center, Sainte-Justine University Hospital Center, 3175, Cote Sainte-Catherine Road, Montreal, Quebec H3T 1C5, Canada.
• Spine Deform. 2018 Jul 1; 6 (4): 351-357.

Study DesignAssessment of vertebral axial rotation measurement methods.ObjectivesTo assess the accuracy and precision of seven radiography-based vertebral axial rotation measurement methods for typical scoliotic deformity before and after posterior instrumentation.Summary Of Background DataVertebral axial rotation is an important component to evaluate transverse plane scoliotic deformities. Several measurement methods were developed based on coronal plane radiographs or computerized 3D reconstruction. Their ability to accurately and precisely measure axial rotation, either pre- or postoperatively, is not well known.MethodsTwo synthetic vertebrae, with and without instrumentation, were fixed in a jig allowing 3D rotation manipulations. Fifty-three configurations of 3D rotations were radiographed. Two observers evaluated seven measurement methods: one visual estimation, two ruler-based (Nash-Moe and Perdriolle), one analytical (Stokes), and three 3D-reconstruction techniques (based on pedicles, based on eight vertebra landmarks, and a surface-based reconstruction software SterEOS). Measurements were repeated one week later.ResultsIntraobserver precision ranged from 2.0° (Perdriolle/SterEOS) to 3.6° (visual estimation) for the noninstrumented vertebra, and from 2.2° (SterEOS) to 9.7° (Nash-Moe) for the instrumented vertebra. Interobserver precision ranged from 1.2° (SterEOS) to 9.3° (Nash-Moe) for the noninstrumented vertebra, and from 1.7° (SterEOS) to 6.2° (Visual Estimation) for the instrumented vertebra. Accuracy of the methods ranged from 2.1° with SterEOS to 9.1° with Nash-Moe ruler. The measurement error was significantly associated with the level of axial rotation for Nash-Moe and 3D reconstruction techniques with low to moderate correlation.ConclusionsThe majority of radiography-based methods measured vertebral axial rotation with an average error of 2° to 5°. The Nash-Moe method should be avoided, considering its inaccuracy greater than 9°. The instrumentation did not compromise the precision or the accuracy of measurement. The measurement accuracy of 3D reconstruction methods was impaired by the severity of the axial rotation.Level Of EvidenceN/A.Copyright © 2017 Scoliosis Research Society. Published by Elsevier Inc. All rights reserved.

15 keywords...

hide…