Despite advances in surgical technique, patellar resurfacing during total knee arthroplasty (TKA) is still relatively unstandardized and carries the risk of patellar maltracking, often linked to intraoperative technical inaccuracies. The aim of this study was to evaluate the relationships between the orientation of the coronal patellar resection plane and resulting patellar tilt angles.

We retrospectively reviewed 303 patients who received cementless patellar resurfacing (2022–2024) during primary robotic assisted TKA. Patients with <6 months follow-up were excluded. Patellar symmetric resection ratio was measured as medial-to-lateral thickness (mm) ratio from the most medial and lateral points on the patellar component backside (1.0 = perfect symmetry) on patella skyline views. Patellar tilt angle measured the amount of patellar tilt in degrees (°) with reference to the anterior trochlea. Patients were stratified into three groups based on patellar tilted resection ratio percentiles: medial tilted resection (<25th percentile [<0.8]), symmetrical resection (25th-75th percentile [0.8−1.1]), and lateral tilted resection (> 75th percentile [>1.1]). This percentile-based approach was chosen to identify clinically interpretable patterns at the distributional extremes while preserving a sufficiently large reference group; continuous regression analyses were retained as the primary statistical framework. ANOVA and Chi-square tests compared between-group differences. Regression assessed the relationship between resection ratio and patellar tilt angle, and identified independent predictors of radiolucent lines surrounding the implant. Additional patient-related variables (e.g., BMI, comorbidities) were not available in the retrospectively collected dataset and could therefore not be entered into the multivariable models.

Mean lateral residual bone thickness was greater than medial (MD = 0.8 mm, P < 0.001) and the mean tilted resection ratio was 0.9 (range, 0.3–2.6). Mean patellar tilt angle was 6.7 ° (range, -11.2–25.7 °). Tilted resection ratio was an independent predictor of patellar tilt (P < 0.001). The medial tilted resection group demonstrated higher mean patellar tilt angle (7.6 ° ± 5.7 °) compared to symmetrical and lateral tilted resection groups (P = 0.022). Radiolucent lines development was independently predicted by patellar tilt angle (OR = 1.1, P = 0.004), but not by tilted resection ratio (P = 0.408). However, medial tilted resection group had the highest radiolucent rate (5.2%), followed by lateral tilted resection (3.9%), and symmetrical resection (2.6%). The optimal patellar tilt threshold for predicting radiolucency was 7.8 ° (sensitivity 90.9%, NPV 99.5%).

A medial tilt of the coronal patella resection plane results in significantly higher risk for patella tilt compared to a symmetric or lateral oblique resection and thus should be avoided. Also, the medial tilted resection group showed the highest rate of radiolucent lines around the pegs, which was again associated with a high patella tilt angle with a cut off at 7.8 °.

III; retrospective comparative study.