Abstract
Background
Recent studies have suggested that full-limb radiographs are more accurate and sensitive than short film radiographs for pre-operative measurement of the anatomical angles required to achieve optimal knee alignment in Total Knee Arthroplasty (TKA). However, there are drawbacks associated with their use including excess radiation to the pelvic organs, the need for specialised radiography equipment and increased cost. Given these drawbacks, we compared the use of MRI scans, a commonly performed pre-operative investigation, with short film knee radiographs for measurement of knee alignment.
Objective
To investigate whether knee alignment measurements made on MRI scans correlate with those measured on short film knee radiographs in patients with osteoarthritic knees.
Methods
We retrospectively reviewed short film knee radiographs and MRI scans of 50 patients with knee osteoarthritis. The plain radiographs had to be performed whilst weight bearing. The MRI scans were performed supine and non-weight bearing. The exclusion criteria included previous trauma to the knee, previous TKA and previous fracture of the lower limb. 4 angle measurements defined by The American Knee Society: alpha, beta, sigma and gamma were measured using each of the 2 modalities. Kolmogorov-Smirnov and two- tailed paired t-tests were used for statistical analysis of the results. Pearson correlation coefficient was used for the measure of dependence.
Results
The alpha, beta, sigma and gamma angles obtained using the MR images were different to those obtained using short film weight bearing knee radiographs by −3°± 1° (p < 0.001), 1° ± 3° (p=0.002), 1° ± 3° (p=0.047) and 1° ± 4° (p=0.113) respectively. There was a weak correlation between the MRI based method and the radiographic method in measuring all 4 angles.
Conclusions
Our results have shown that the angular measurements performed on MR images should be interpreted with caution as they may vary depending on the MRI slice selected for evaluation. The differences observed and the weak correlation between the 2 modalities may be due to the different scopes used for determining the femoral or tibial axes. The measurements obtained using the plain radiographs were interpreted using a single 2D projection of a 3D structure. However, the measurements of the MR images were evaluated using a 2D image of a slice through a 3D structure. In conclusion, the use of MRI scans for pre-operative planning in TKA may not be advisable at this stage as the angular measurements obtained using the MR images were poorly correlated to those obtained using plain radiographs.
