Coronal Plane Alignment of the Knee Classification Remains Best Determined on Long-Leg Radiographs Compared to Computed Tomography Scanograms or Short Knee Radiographs.

Abstract

The coronal plane alignment of the knee (CPAK) classification is a pragmatic method to characterize constitutional limb alignment. The CPAK classification was originally described using radiographic measurements from standing long-leg radiographs (LLRs). Surgeons who do not obtain LLRs before total knee arthroplasty (TKA) may estimate CPAK using alternative modalities, including supine computed tomography (CT) scanograms and standing short knee radiographs. This study was performed to evaluate the correlation in measurements of lower limb alignment and the agreement of CPAK classification between these three imaging modalities. A retrospective review of 300 knees undergoing robotic-assisted primary TKA was performed. All patients had preoperative LLRs, CT scanograms, and short knee radiographs. There were two reviewers who independently measured the mechanical lateral distal femoral angle (mLDFA) and mechanical medial proximal tibial angle on each modality. A five-degree adjustment was made to the anatomical lateral distal femoral angle measured on short knee radiographs to calculate an estimated mLDFA. Correlation coefficients for measurements between imaging modalities were evaluated. The agreement of CPAK phenotypes using each modality was assessed. When compared to LLR measurements, strong correlations were found for the mLDFA and mechanical medial proximal tibial anglemeasurements from CT scanograms (correlation coefficient, 0.839 and 0.819, respectively) and short knee radiographs (0.845 and 0.896, respectively). Moderate agreement for the CPAK type on LLRs was determined from CT scanograms (61.7% agreement; Cohen's kappa, 0.455) and short knee radiographs (68.7% agreement; 0.571). However, a different CPAK classification versus LLRs was determined from CT scanograms and short knee radiographs in over 30% of knees. Despite a strong correlation between imaging modalities for all radiographic measurements of limb alignment, only moderate agreement was noted in CPAK types. Therefore, surgeons who use CPAK classification to determine surgical alignment strategy in TKA, or when reporting outcomes associated with CPAK, should use LLRs.