Three-dimensional joint surface orientation does not correlate with two-dimensional coronal joint line orientation in knee osteoarthritis: Three-dimensional analysis of upright computed tomography

Background: Two-dimensional (2D) coronal joint line orientation on radiography under weight-bearing conditions has been widely used in evaluating knee osteoarthritis (OA). However, the effects of tibial rotation remain unknown. The present study aimed to newly define three-dimensional (3D) joint surface orientation relative to the floor as an unchangeable 3D angle unaffected by tibial rotation using upright computed tomography (CT), and to investigate correlations between 3D and 2D parameters in knee OA. Methods: Sixty-six knees in 38 patients with varus knee OA underwent standing hip-to-ankle digital radiography and upright CT. The 2D parameters measured on radiographs included femorotibial angle (FTA), tibial joint line angle (TJLA), lateral distal femoral angle (LDFA), medial proximal tibial angle (MPTA), and joint line convergence angle (JLCA). The 3D inner product angle between vectors of the tibial joint surface and the floor from CT was defined as the 3D joint surface-floor angle. Results: Mean 3D joint surface-floor angle was 6.0 ± 3.6°. No correlation was identified between 3D joint surface-floor angle and 2D joint line parameters, even though FTA correlated substantially with 2D joint line parameters. No significant differences in 3D joint surface-floor angle were identified among Coronal Plane Alignment of the Knee (CPAK) types. Conclusions: The 3D joint surface orientation did not correlate with 2D coronal joint line orientation and was unaffected by CPAK classification types. This finding suggests that current 2D evaluations should be reconsidered for a better understanding of true knee joint line orientation.