Comparison of CT and EOS in assessing coronal lower limb alignment when planning total knee arthroplasty

Background: As surgical technologies and alignment strategies develop, accuracy of lower limb alignment assessment gains increasing importance. The current gold standard remains long leg (4ft) radiographs. Other measures include computed tomography (CT) and EOS scans. This study aimed to compare CT and EOS long leg views to determine the reliability of assessment of hip-knee-angle (HKA) in arthritic knees.

Materials and methods: A retrospective study of 96 knees in patients undergoing total knee arthroplasty (TKA) was performed comparing HKA alignment data from EOS and CT. Coronal HKA and sagittal flexion angle were assessed by two independent observers. Inter-observer correlation was calculated.

Results: The coronal alignment on EOS was strongly positively correlated to coronal alignment on CT (r = 0.927, p = 0.001). 12 knees (13 %) exceeded CT vs EOS difference of 3°. Inter-rater reliability was excellent (intra-class coefficients >0.9). The mean difference between CT and EOS was significantly greater for patients with fixed flexion >10° (0.68) vs <10° (-0.2) p = 0.004. Mean difference in HKA did not differ between those 0-10° varus and >10° varus (p = 0.273). Valgus HKA had a higher mean difference (1.9°) compared to varus knees (-0.4°) (p = 0.001).

Conclusion: CT and EOS showed excellent inter-rater reliability and correlated well. Increased sagittal plane deformity does effect coronal HKA assessment. Extreme varus did not affect the mean difference significantly, while valgus did. For the majority of patients either CT or EOS will give a reliable assessment of HKA but beware those with significant valgus or sagittal deformity where both modalities may be necessary to plan TKA.