Abstract
Introduction
Loads acting on the knee are tied to the long term performance of implants, and are directly related to ligament function [1]. Previous work has used computational models coupled with optimization to estimate ligament properties based on experimental joint kinematics [2]. Our group recently utilized a similar optimization scheme that estimated ligament slack lengths based on experimental implant contact metrics [3]. A comparison with surgically relevant loading conditions that were excluded from the optimization would help establish the utility of the simulation framework. Hence, the purpose of this study was to assess the predictive capability of two simulated knees using comparisons with experimentally determined trends found after systematic removal of key tissues. Similar techniques may support clinical joint balancing techniques, as well as identify factors that dictate long term implant performance.
Methods
Knee arthroplasty was performed by orthopedic surgeons for four cadaveric specimens. Instrumented trial inserts (VERASENSE, OrthoSensor, Inc., Dania Beach, FL) were used and experimentation utilized the simVITROTM robotic musculoskeletal simulator (Cleveland Clinic, Cleveland, OH) to measure tibiofemoral kinematics under interoperative style loading. Three successive laxity style tests were performed at 10° flexion: anterior-posterior force (±100 N), varus-valgus moment (±5 Nm), and internal-external moment (±3 Nm). Kinematics and implant forces were measured throughout testing. Specimens were first tested in the intact state, then the laxity tests were repeated after systematic release of the posterior cruciate ligament (PCL), superficial medial collateral ligament (sMCL), or popliteus (POP). Significant changes in kinematics and contact metrics were determined using regression analysis between the intact versus the tissue released states.
Finite element models were developed for two specimens, and optimized ligament slack lengths were found using methods described previously [3] (Fig. 1). The experimental laxity style loads were applied to both optimized models with intact ligaments, and with individually released PCL, sMCL, or POP ligaments. Knee kinematics and tibial contact loads were predicted, and trended responses from the intact simulations to those with released ligaments were determined (i.e. higher, lower or no change). Simulation results were then compared with the statistically significant findings from the experimental tests.
Results and Discussion
Both models generally recreated the significant experimental trends. Specimen 3 recreated 8 of the 9 directional changes, while specimen 4 realized 7 of the 9 (Table 1). Release of the POP in specimen 4 contradicted both specimen 3 and the experimental results. This may highlight specimen-specific behavior, or a misrepresentation of the tissue restraint on the posterolateral corner. Ongoing testing and simulation will evaluate areas of discrepancy, with particular focus on specimen specific mechanics.
This work shows that simulation can estimate significant trends in physical testing. The framework demonstrates promise for development of a tool to understand the consequences of intra-operative tissue balancing. Future work will investigate representation of the posterolateral corner, and evaluate the predictive capacity for the absolute specimen-specific changes in joint mechanics due to tissue release.
Acknowledgements
Orthosensor Inc.
For any figures or tables, please contact authors directly (see Info & Metrics tab above).
