The Influence of Modeling Choices on the Predictive Capability of the Human Knee Joint Finite Element Model

Objective

In finite element knee models ligaments have been represented either by a group of one-dimensional springs, or by

three-dimensional continuum elements based on segmentations. Moreover, the knee

ligaments mechanical properties might be either assigned from the reported data in the literature, or adjusted specifically for the subject.

the implications of modeling strategies on the resulting joint biomechanics are currently unknown.

The aim of this study was to evaluate the effects of:

a) The ligament modeling approach

(non-linear springs (1D) vs. transversely isotropic continuum (3D) models); and

b) The selection of the data used to describe the behavior of ligaments (based either on the literature, or on subject-specific

optimized values),

on biomechanical predictive abilities of FE models of human native knee joints.

The Influence of Modeling Choices on the Predictive Capability of the

Human Knee Joint Finite Element Model

Hamid Naghibi Beidokhti

_{, Dennis Janssen}

_{, Sebastiaan van de Groes}

_{, Ton van den Boogaard}

_{, Nico Verdonschot}

1 _{Orthopaedic Research Lab, Radboud University Nijmegen Medical Centre, The Netherlands,} 2 _{Orthopaedic Department, Radboudumc, Nijmegen, The}

Netherlands, 3 _{Department of Applied Mechanics, University of Twente, The Netherlands,} 4 _{Department of Biomechanics, University of Twente, The}

Netherlands.

Hamid Naghibi, MSc

Hamid.NaghibiBeidokhti@radboudumc.nl

Orthopaedic Research Laboratory Radboud university medical center

P.O. Box 9101, 6500 HB Nijmegen, The Netherlands www.biomechanics.nl

This study is a part of BIOMECHTOOLS funded by European Research Council under the Seventh Framework Program

(FP/2007-2013) / ERC Grant Agreement n. 323091

Methods

In-vitro experiments were performed on

three human cadaver knee joints with a 6-DOF knee testing apparatus (Fig.1). For each specimen, two FE models were

developed with ligaments modeled either using springs, or using continuum

representations in Abaqus v6.13 (Pawtucket, RI, USA).

A series of laxity tests were applied based on which stiffness parameters and pre-strains were optimized for both modeling

approaches using Isight (Simulia, Providence, RI).

Validation experiments were conducted to evaluate the biomechanical outcomes of the FE models with the two ligament

modeling techniques, with either literature-based or subject-specific

parameters.

Conclusions

in FE, adopting subject-specific material parameters considerably

affects and improves the quality of the model predictions.

o Using a continuum modeling approach results in more accurate contact

outcome variables.

o However, when mainly the prediction of joint kinematics is of interest, the spring ligament models provide a faster option.

Results

Models (both spring and continuum) with subject-specific properties improved the predicted kinematics and contact

outcome parameters. Models based on literature-based parameters, and

particularly the spring models, led to relatively high errors in kinematics and contact pressures (Fig. 2 & Table1).

Average RMS difference ± standard deviation

Unloaded deep flexion Tibia axially loaded deep flexion Tibia anteriorly loaded flexion

Li te ra ture -base d spring m odel Optimize d spring m odel Li te ra ture -base d cont inuum m odel Optimize d cont inuum m odel Li te ra ture -base d spring m odel Optimize d spring m odel Li te ra ture -base d cont inuum m odel Optimize d cont inuum m odel Li te ra ture -base d spring m odel Optimize d spring m odel Li te ra ture -base d cont inuum m odel Optimize d cont inuum m odel Internal/External rotations in ° 5.7±2.1 2.8±1.1 2.5±1.2 1.8±0.2 5.6±2.7 3.0±0.4 3.1±1.0 2.2±1.2 6.7±3.4 2.6±1.1 4.1±1.1 3.8±0.8 Varus/Valgus rotations in ° 5.6±2.0 3.5±0.3 5.0±2.6 2.9±0.9 5.7±1.8 3.8±0.6 3.9±1.4 2.7±1.0 4.1±2.4 3.2±1.7 4.1±0.5 4.3±1.3 Anterior/Posterior translations (mm) 5.8±2.1 3.0±2.2 7.4±2.2 3.8±1.5 5.8±2.3 2.9±2.2 7.4±2.8 4.1±1.4 3.1±1.3 2.2±0.4 2.9±1.3 2.6±1.6 Lateral/Medial translations (mm) 2.5±2.0 1.6±1.7 2.1±1.1 1.9±1.2 3.4±1.4 3.1±1.7 3.4±2.5 2.4±1.6 3.1±1.4 2.8±0.8 2.7±1.1 2.8±1.1

Figure 2: Contact pressure at tibial cartilages (flexion:90°) for all three specimens from

experiment and with different modeling approaches

Figure 1: The illustration of the current study methodology

Table 1: Average RMS (three specimens) difference between experimental validation tests kinematics and different ligament modeling approaches.

View publication stats View publication stats