Osteoarthritis (OA) is one of the leading causes of disability in the United States, with total knee arthroplasty (TKA) a common end stage treatment for knee OA. While generally successful at relieving pain, some TKA patients have suboptimal outcomes including walking and climbing stairs more slowly than control groups, and difficulty squatting. Surgical technique, such as prosthetic component alignment and soft-tissue balancing, has been identified as critical to success of a TKA, but definitive links that would allow for predictions of post-operative outcomes based on intra-operative variables have yet to be established. Surgical navigation systems have been developed to reduce errors in component alignment and provide some measures of intra-operative variables, but clinical benefits are unclear, and they are unable to make predictions of post-operative outcomes. Computer simulations have been used to establish general trends between technique and post-operative outcomes. However, simulations may need to be patient-specific in order to be used in a predictive fashion due to inter-patient variations in parameters, such as ligament properties, that directly affect simulation results. While some studies have estimated patient-specific properties in cadaveric specimens, a methodology for estimating properties in TKA patients has yet to be presented. It is also unknown how patient-specific properties may impact simulations of post-operative function such as gait.

We introduced a methodology for estimating knee soft-tissue properties of an individual total knee patient using a custom surgical navigation system and stability device. The force-displacement relationship of the knee was measured, and soft-tissue properties were estimated using a parameter optimization that matched simulated tibiofemoral kinematics with experimental tibiofemoral kinematics. Simulations with optimized properties more closely match experimental data than simulations with generic properties (RMS error of 3.5° for optimized and 8.4° for generic), while specimens showed large variability among ligament properties (e.g. range of 2600N for stiffness of the superficial medial collateral ligament) regardless of similarities in prosthetic component alignment and measured knee laxity. These results demonstrate the importance of soft-tissue properties in determining knee stability, and suggest that to make clinically relevant predictions of post-operative knee motions and forces using computer simulations, patient-specific soft-tissue properties are needed.

To extend the methodology to include patients within the operating room, we collected intra-operative knee stability data and used them to estimate the knee ligament properties of three patients. We then compared a simulation’s ability to predict experimental passive range of motion room using the optimized ligament properties as well as generic ligament properties taken from literature. We found that RMS errors for the knee stability tests were similar for both optimized and generic simulations as were found in our previous cadaver study, and that optimized simulations showed an improvement in predictions of passive range of motion tests compared with generic simulations. This suggests that using patient-specific properties may play an important role in accurately predicting intra-operative tests of passive range of motion.

Lastly, we performed forward simulations of TKA gait to investigate the impact of patient-specific ligament properties on simulations of post-operative functions. We developed a modified dual-joint methodology that took advantage of OpenSim API functions and implemented them using custom Matlab code and built-in Matlab integrators. Utilizing the dual joint methodology allowed us to calculate the tibiofemoral and patellofemoral contact forces in the TKA knee. Results from these simulations showed improved tracking of gait kinematics using patient-specific properties in the simulation. Contact forces and muscle forces were affected by choice of ligament properties, with differences seen throughout the gait cycle.

This dissertation advances our understanding of the role that choices in ligament properties make in simulations of the TKA knee, and introduces a novel methodology for estimating the ligament properties of TKA patients based on data collected intraoperatively. These studies lay the groundwork for future research to develop tools that aid surgeons in developing an optimal surgical plan, and which ultimately lead to improvements in quality of life and surgical outcomes for those undergoing a TKA.