Three-Dimensional Computed Tomography Reconstructions Do Not Improve the Interobserver Reliability of the AO/Orthopaedic Trauma Association and Young and Burgess Pelvic Fracture Classifications.

Abstract

To evaluate the value of three-dimensional (3D) computed tomography (CT) scans on the interobserver and intraobserver reliability of AO/Orthopaedic Trauma Association (OTA) and Young and Burgess (YB) classifications for pelvic ring injuries. Seven reviewers (four fellowship-trained orthopaedic trauma surgeons and three fellows) independently classified 36 pelvic ring injuries using radiographs and axial two-dimensional (2D) CT scan images and then repeated this process 2months later with the addition of 3D CT images. Interobserver and intraobserver reliability was assessed. The interobserver reliability of the AO/OTA classification using 2D vs. 3D CT scans was considered fair (k 0.23, CI 0.17-0.29) vs. slight (k 0.16, CI 0.09-0.22), with no observed difference [mean difference (MD) -0.07, CI 0.16-0.01]. The interobserver reliability of the YB classification using 2D vs. 3D CT scans was considered fair for both (k 0.37, CI 0.32-0.42, vs. 0.37, CI 0.30-0.45), with no observed difference (MD -0.0005, CI -0.08 to 0.08).The intraobserver reliability of the AO/OTA vs. YB classifications was considered fair (k 0.35, CI 0.26-0.44) vs. moderate (k 0.49, CI 0.40-0.57), with the YB classification having higher kappa value (MD 0.13, CI 0.01-0.26, p = 0.03). The addition of 3D CT scan reconstructions to radiographs and 2D CT did not improve the interobserver reliability of AO/OTA and YB classifications for pelvic ring injuries.