Design phase evaluation of potential implant designs requires verified computational and experimental models. Computational models are important where parametric evaluation of geometric features experimentally is both cost and time-prohibitive due to the need to manufacture complex parts, and provide information not easily measured experimentally, such as internal stresses/strains in the implant or bone. However, before implementation into the design process, a thorough verification/validation is required. In this study, a finite element model of the Kansas knee simulator (KKS) was developed and a systematic verification of predicted joint kinematics was performed by comparison with experimental measurements, including evaluating the patellofemoral joint first in isolation, followed by whole joint kinematic comparisons. Four unmatched, healthy cadaver knees (average age 63 yrs) were mounted in the KKS to reproduce a simulated gait and deep knee bend activity in their natural and implanted states. Finite element models of the KKS assembly and the four cadaver specimens in their natural and implanted states were created. Isolated patellofem-oral kinematics were initially verified during simulated deep knee bend. Average RMS differences between predicted and experimental natural patellar kinematics were less than 3.1° and 1.7 mm for rotations and translations, respectively, while differences in implanted kinematics were less than 2.1° and 1.6 mm between 10 and 110° femoral flexion. Similar agreement was found with the subsequent whole joint simulations. Deep knee bend tibiofemoral internal-external (IE) and varus-valgus (VV) rotations had average RMS differences from experimental measurements of 1.5 ± 0.4° and 0.9 ± 0.5°, respectively. Anterior-posterior (AP), inferior-superior (IS) and medial-lateral translations matched within 1.8 ±0.8 mm, 1.2 ±0.7 mm, and 0.6 ±0.1 mm, respectively. The experimental and verified computational tools can be used in harmony for pre-clinical assessment of implant designs; the computational model allows rapid screening of implant geometry or alignment issues and provides additional insight into joint mechanics such as implant stresses or bone strains, while the experimental simulator can subsequently be utilized to assess in cadavera only the most promising designs or features identified.
Patient charts and radiographs were reviewed. Statistical analysis was performed. Significant variables associated with patient anatomy, implant size and alignment were subsequently investigated in a computational model to evaluate tendofemoral contact.
