Abstract
Introduction
While fluoroscopic techniques have been widely utilized to study in vivo kinematic behavior of total knee arthroplasties, determination of the contact forces of large population sizes has proven a challenge to the biomedical engineering community. This investigation utilizes computational modeling to predict these forces and validates these with independent telemetric data for multiple patients, implants, and activities.
Methods
Two patients with telemetric implants, the first of which was studied twice with the reexamination occurring 8 years after the first, were studied. Three-dimensional models of the patients' bones were segmented from CT and aligned with the design models of the telemetric implants. Fluoroscopy was collected for gait, deep knee bend, chair rise, and stair activities while being synchronized to the ground reaction force (GRF) plate, telemetric forces, knee flexion angles, electromyography (EMG), and vibration sensors. Registration of the implants and bones to the 2-D fluoroscopy provided the 6 degree of freedom kinematic data for each object. Orientation and position of the components, the GRFs, ligament properties, and muscle attachment locations were the only inputs to the Kane's dynamics inverse solution. Dynamic contact mapping and pseudo-inverse solution method were incorporated to output the predicted muscle forces of the vastus lateralis, rectus femoris, vastus medialis, biceps femoris long head, and gastrocnemius and contact forces at the patellofemoral and medial and lateral tibiofemoral. While every major muscle of the lower limb was incorporated into the model, these five were used in the validation process. EMG signals were processed to determine the neural excitation, muscle activation, and using the dynamic muscle length from the kinematics, the tension generated by these muscles.
Results
Comparison of the model predictions for the tibiofemoral contact forces with the telemetric implant data resulted in an error <10% for all patients and activities. Predicted muscle forces were <15% error from the EMG calculated forces.
Discussion
An inverse computational model of the knee robust enough to encompass multiple patients and activities was successfully created and validated. The accuracy of the muscle forces demonstrates that the model correctly simulates anatomical motion and not just transferal of GRFs. While this study was conducted on patients with telemetric implants, the required inputs to the model can be obtained from any TKA patient with the mobility to conduct the desired activity. This allows not only kinematic data, but also kinetics, to be provided for the improvement of implant design and surgical techniques accessibly and relatively inexpensively.
