CAD-Based Analysis and Experimental Validation of Registration Errors in Imageless Total Knee Arthroplasty

Background/Objectives: Accurate implant positioning in total knee arthroplasty (TKA) depends on reliable intraoperative landmark registration. In imageless TKA, registration errors can alter cutting-plane orientation and compromise alignment. This study quantitatively evaluated how anatomical landmark registration errors affect cutting-plane orientation in imageless TKA. Methods: A CAD-based simulation with controlled experimental validation using 3D-printed bone models was performed to reproduce the imageless TKA workflow. Controlled errors were introduced into key femoral and tibial landmarks, and the resulting deviations were quantified. The primary evaluation metrics were angular deviations in varus/valgus, flexion/extension, and internal/external rotation. Results: Coronal and rotational alignment showed the greatest sensitivity to registration error. In the femur, anteroposterior epicondylar displacement had the strongest rotational influence, with sensitivity reaching about 0.5°/mm, whereas mediolateral displacement of the tibial anteroposterior landmarks showed the highest sensitivity at about 1.4°/mm. Similar trends were observed in both simulation and experimental validation cases. Conclusions: The findings indicate that small registration errors can produce clinically significant cutting-plane deviations in imageless TKA, particularly at the femoral transepicondylar and tibial anteroposterior landmarks, and may approach commonly accepted alignment thresholds.