A modular, mechanical knee model for the development and validation of robotic testing methodologies.

Abstract

Six-degree-of-freedom robotic testing is used to gain insight into knee function by measuring the biomechanics of cadaveric knees. However, it can be challenging to use cadaveric knees to validate robotic testing methodologies and to compare methodologies across laboratories because cadavers have variable properties and require lengthy preparation. Therefore, our primary objective was to develop a modular, mechanical knee model for robotic testing with comparable biomechanics to those of human cadaveric knees. A secondary objective was to use the knee model to benchmark the errors in ligament tensions measured using the superposition method, which is a common robotic testing workflow to determine in situ ligament tensions. We designed a knee model consisting of femur and tibia components that are constrained by their articular geometries and by ligament phantoms. We used our robotic testing system to measure the kinetic-kinematic relationships under anterior-posterior, varus-valgus, and internal-external loading in four knee models with different design features. We achieved variable kinetic-kinematic relationships across the knee models by adding secondary restraints, altering the engagement of the ligament phantoms, and incorporating osteoarthritic features. The knee models had comparable laxities to cadaveric knees, although most knee models did not capture the flexion-dependent kinematics of cadaveric knees. We also found comparable errors in superposition-computed tensions in the lateral collateral ligament between the knee models and cadaveric knees. Therefore, the knee model is a biomechanically representative specimen that can be a valuable tool for developing and validating robotic testing methodologies.