An Efficient Patient-Specific Biomechanical Model of Total Knee Arthroplasty and Its Validation Based on In Vivo Force Measurements

Introduction: In total knee arthroplasty the individual bone morphology is changed by the artificial implant geometry and placement, which might induce incompatibilities in the sensitive joint mechanisms, causing patient dissatisfaction and premature implant failure. Therefore, a comprehensive understanding of the governing parameters and a study of their effects on biomechanics of the individual patient are essential. The goal of our study was to develop and validate a simplified biomechanical simulation model considering articular surfaces where the patient-specific adaption process was aimed to be valid for data, which are commonly available in the clinical workflow.

Methods: The experimental data used in this study were part of the “Grand Challenge Competition to Predict In Vivo Knee Loads” (GCC). For the model adaption and validation the data from GCC3, GCC5 and GCC6 were used, since these were complete and the most reliable. The patient-specific model was developed from scratch in the software AnyBody. The lower extremity is represented by 3 segments (femur, tibia, patella) and the tibiofemoral joint contains 6 and the patellofemoral joint 5 degrees of freedom. The basic flexor and extensor muscles activated during two-leg squat, which has broad consensus to several everyday activities, were included. In order to stabilize the knee joint 6 ligamentous structures were included representing the posterior cruciate, collateral and patellofemoral ligaments. A generic pelvis served as body dummy to apply an external force to the lower extremity. The antagonistic ground reaction force was directly estimated based on the Newton–Euler equations of motion. The medial and lateral knee contact forces were recorded during two-leg squat simulation and compared to the corresponding in vivo measurements.

Results: The lateral contact forces compared well with the in vivo forces of all 3 patients (RMSE < 21 %BW) and the medial forces for GCC3 and GCC5 (RMSE < 26 %BW). However, for GCC6 the medial force was overpredicted which resulted in an overall high total force. Nevertheless, the differences of the total forces for GCC3 and GCC5 were small (RMSE < 26 %BW).

Discussion: The presented validated biomechanical model offers the opportunity to predict and study mechanics and kinematics which has the potential for integration into clinical workflow and to improve implant design and surgical outcome after TKA.