CT-based CPAK measurements show more varus MPTA and valgus LDFA than LLRCT Scans Show Different Knee Alignment Than X-rays

INTRODUCTION: Coronal plane alignment of the knee (CPAK) parameters are increasingly used to characterize native coronal alignment and guide alignment strategies in total knee arthroplasty (TKA) based on long-leg radiographs (LLR). Computed tomography (CT) based robotic TKA software enables calculation of CPAK parameters based on anatomical reference points. However, potential differences in CPAK measurements between LLR and CT remain incompletely defined. This study compared CPAK coronal alignment parameters and phenotype distributions between LLR and CT, and evaluated factors associated with measurement discrepancies. METHODS: This was a radiographic subanalysis of a prospective randomised controlled trial including 241 patients undergoing robotic-assisted TKA. Medial proximal tibial angle (MPTA) and lateral distal femoral angle (LDFA) were measured and arithmetic hip-knee-ankle-angle (aHKA), joint line obliquity (JLO) and CPAK phenotype distribution were calculated and classified from preoperative LLR and CT. Discrepancy groups were defined as LLR = CT (≤ 2°), LLR < CT (> 2°) and LLR > CT (> 2°). Associations between discrepancy groups and preoperative deformity and flexion contracture were assessed. RESULTS: Compared with LLR, CT demonstrated a more varus MPTA (mean 86.7° vs. 87.1°, p = 0.04), a more valgus LDFA (87.1° vs. 87.8°, p < 0.01) and a lower JLO (173.9° vs. 174.9°, p < 0.01). The aHKA did not differ between modalities. There was a higher overall proportion of apex distal phenotypes (Types I-III) on CT (p = 0.01), and in particular CPAK Type II phenotypes were more frequent on CT than LLR (45.2% vs. 33.2%; p < 0.01). For MPTA, the proportions of LLR = CT, LLR > CT, and LLR < CT were 66.0%, 21.6% and 12.4%, respectively; corresponding LDFA proportions were 84.6%, 13.3% and 2.1%. In MPTA discrepancy, LLR > CT groups showed more preoperative valgus alignment and LLR < CT group showed more preoperative varus alignment with knee flexion contracture. CONCLUSION: CPAK coronal alignment parameters and phenotype distributions differ depending on whether measurements are derived from LLR or CT. Discrepancies in MPTA are influenced by preoperative coronal knee alignment and flexion contracture. Surgeons should interpret CT- and LLR-based CPAK assessments with caution, particularly in patients with moderate deformity or flexion contracture. LEVEL III: Retrospective Cohort study.