Abstract
Introduction
Radiographs and computed tomography (CT) images are used for the preoperative planning in total knee arthroplasty (TKA), however, these two-dimensional (2D) measurements are affected easily by limb position and scanning direction relative to three-dimensional (3D) bone model analyses. The purpose of our study was to compare these measurements to evaluate the factors affecting the difference.
Patients and Methods
A total of 75 osteoarthritis knees before primary TKA were assessed. The full-length weight-bearing anteroposterior radiograph and CT slices were used for the 2D measurement. Three-dimensional measurement used 3D bone model reconstructed from the CT data and the coordinate system as the previous reports (Figure 1). We measured FVA (femoral valgus angle), CRA (the angle between the posterior condylar line <PC-L> and the clinical epicondylar axis <CEA>), and SRA (the angle between the PC-L and the surgical epicondylar axis <SEA>). Intra- and inter-observer reliabilities were assessed by intraclass correlation coefficients (ICC), and the differences between the 2D and the 3D measurements (Differences) were evaluated. In addition, we evaluated whether preoperative factors (preoperative extension angle, HKA, BMI and CT scanning direction) affected the differences between the 3D and the 2D measurements. Computer simulation was used to examine the influences of CT scanning direction.
Results
The ICC and the mean values with 2D and 3D measurements were shown in Table 1. The mean Differences were 0.2 ± 1.2° in FVA, 0.7 ± 2.1° in CRA and 0.7 ± 1.8° in SRA. Nine percentage in CRA and 13% in SRA had over 3 degrees of the Differences. There were no significant correlations between these factors and the Difference, however, the 3D simulation showed a significant difference of CRA between the scanning direction in varus/valgus and the neutral condition (varus: p<0.001, valgus: p<0.001) (Figure 2).
Discussion
From our study, the 3D measurements were highly reliable. In the 2D measurements, the rotational measurements showed less interobserver reliability because of the difficulty of picking the same CT plane and the same position of femoral bony landmarks between observers. The mean Differences were small, however, the range was large and almost10% of our cases were over 3 degrees which can induce the malalignment of the component in spite of the precise bone cutting. Particularly, smaller values of the CRA and SRA with the 2D measurement have a risk of internal-rotated position of the femoral component. Preoperative osteoarthritis knees have flexion contractures, valgus, and femoral torsion. The 3D measurements are useful to avoid the different evaluation from the CT scanning situation.
