Two-dimensional measurement misidentifies alignment outliers in total knee arthroplasty: a comparison of two- and three-dimensional measurements

Purpose Two-dimensional (2D) and three-dimensional (3D) measurements of prosthetic alignment and “outliers” after total knee arthroplasty (TKA) might not necessarily be comparable. The aim of this study was to compare the use of 2D and 3D measurements in the identification of prosthetic alignment and outliers after TKA. Methods This cross-sectional study included 159 consecutive TKAs. All patients underwent plain radiography, fluoroscopy-guided radiography, and computed tomography after TKA. The same baseline was used for 2D and 3D measurements. The reliability of prosthetic alignment and outlier identification (> 3° from neutral alignment) was compared between the 2D and 3D measurements. Results The mean prosthetic alignment and rate of outliers were not significantly different the between 2D and 3D measurements. The inter- and intra-observer reliabilities were higher for the 3D measurements than for the 2D measurements. The agreement between 2D and 3D measurements in outlier identification was poor, except for femoral coronal alignment. Cohen’s κ coefficients were 0.19 in femoral sagittal (poor), 0.02 in tibial coronal (poor), and 0.10 in tibial sagittal (poor) on plain radiography. The values were 0.23 in tibial coronal (poor) and 0.002 in tibial sagittal (poor) on fluoroscopy-guided radiography. Conclusions The agreement between 2D and 3D measurements in the identification of outliers was poor, even on fluoroscopy-guided radiography. Since 2D measurements have an inherent risk of misidentifying alignment outliers, this finding is clinically relevant. To properly analyze the correlation between outliers and clinical results, such as longevity, patient satisfaction, and patient-reported outcome, 3D measurements for prosthetic alignment are desirable. Level of evidence III.