A comparison of femoral component rotation after total knee arthroplasty in Kanekasu radiographs, axial CT slices and 3D reconstructed images

Objective To compare the posterior condylar angle measured with Kanekasu radiograph and 2D-CT with the gold standard 3D-CT following primary total knee arthroplasty (TKA). Methods Eighty-two knees with pain following TKA were included in this retrospective study. Two independent raters measured the anatomical and surgical posterior condylar angles twice on each Kanekasu radiograph and 2D-CT. These measurements were compared against the 3D-CT measurement. The intra- and interrater reliability of the Kanekasu radiograph and 2D-CT and the correlation with 3D-CT were calculated. Results The intra- and interrater reliability for measurements of the anatomical posterior condyle angle for the Kanekasu radiograph and the 2D-CT were excellent for both raters (0.85-0.92). For the less experienced rater 1, the intrarater reliability was significantly better for 2D-CT than Kanekasu radiograph for measuring both the surgical (p < 0.01) and anatomical posterior condyle angles (p < 0.05). For the experienced rater 2, the intrarater reliability was significantly better for Kanekasu radiograph than 2D-CT for measurement of the surgical posterior condyle angle (p < 0.05). The correlation with 3D-CT is higher in 2D-CT than in Kanekasu radiograph (p < 0.01). While the Kanekasu radiograph predicts the 3D-CT angle with 65.9%, 2D-CT can measure the true angle with 82.9% certainty. Conclusion Measurements using the anatomical transepicondylar axis are easier to replicate compared to the surgical transepicondylar axis. In comparison with the gold standard 3D-CT, 2D-CT showed a significantly higher correlation with 3D-CT than the Kanekasu measurements. If 3D-CT is available, it should be preferred over 2D-CT and Kanekasu view radiograph for femoral component rotation measurements.