Inter- and Intraobserver error in distal femur transepicondylar axis measurement with computed tomography

Abstract

INTRODUCTION: Correct rotational alignment of the femoral component in total knee replacement is essential for satisfactory patellar tracking. Measurement of the rotation of the femoral components by computed tomography (CT) with reference to the transepicondylar axis (TEA) is regarded as the gold standard in most reports. However, the paucity of evidence validating the accuracy of it is surprising. METHODS: On 2 separate occasions at least 1 week apart, 6 independent observers (4 orthopaedic surgeons and 2 radiologists) studied the inter- and intraobserver error in identification of the TEA using CT in 10 cadaveric knees. The angle between TEA and posterior condylar line was measured in both conventional two-dimensional axial CT images and three-dimensional CT reconstruction images. The reference surgical TEA was established by dissection of specimens. RESULTS: We found an average error of 2.6° (external rotation) when the TEA was identified by CT. The error was significantly smaller when it was identified by using a conventional axial image than when a three-dimensional reconstruction image was used (P<0.001; paired t-test). No significant intraobserver error (P=0.814; Wilcoxon rank test) was found when the measurement was performed using conventional two-dimensional CT images. However, significant interobserver difference was evident (P<0.001; Friedman test). DISCUSSION AND CONCLUSION: The average error in identification of the TEA by CT is 2.6° ex ternal rotation. The use of three-dimensional reconstruction images did not provide any additional benefit over the two-dimensional CT measurement; we observed significant interobserver differences with it.