Rotational Projection Errors in Coronal Knee Alignment on Weight-Bearing Whole-Leg Radiographs: A 3D CT Reference...

Whole-leg radiographs (WLRs) are widely used to assess coronal alignment before total knee arthroplasty (TKA), but may be inaccurate in patients with atypical morphotypes or malrotation. This study evaluated the discrepancy between WLR and 3D computed tomography (CT) scans across coronal plane alignment of the knee (CPAK) morphotypes and introduced a novel projection index-the femoral notch projection ratio (FNPR). In CPAK III knees, 19% of cases exceeded a clinically relevant threshold (>3° difference), prompting investigation of underlying projection factors. In 187 knees, coronal angles-including the medial distal femoral angle (MDFA°), medial proximal tibial angle (MPTA°), femoral mechanical angle (FMA°), and arithmetic hip-knee-ankle angle (aHKA°)-were measured using WLR and CT. Rotational positioning on WLR was assessed using FNPR and the patellar projection ratio (PPR). CPAK classification was applied. WLR systematically underestimated alignment, with the greatest bias in CPAK III (MDFA° + 1.5° ± 2.0°, p < 0.001). FNPR was significantly higher in CPAK III and VI (+1.9° vs. -0.3°, p < 0.001), indicating a tendency toward internally rotated limb positioning during imaging. The PPR-FNPR mismatch peaked in CPAK III (4.1°, p < 0.001), suggesting patellar-based centering may mask rotational malprojection. Projection artifacts from anterior osteophytes contributed to outlier measurements but were correctable. Valgus morphotypes with oblique joint lines (CPAK III) were especially prone to projection error. FNPR more accurately reflected rotational malposition than PPR in morphotypes prone to patellar subluxation. A 3D method (e.g., CT) or repeated imaging may be considered in CPAK III to improve surgical planning.