Development of a three-dimensional finite element model of a human mandible containing endosseous dental implants. II. Variables affecting the predictive behavior of a finite element model of a human mandible.

The purpose of this study was to propose a systematic approach to validate a finite element model (FEM) of the human mandible and to investigate the effects of changing the geometry and orthotropic material properties on the FEM predictions. Thirty-eight variables affecting the material properties, boundary conditions, and the geometry of a FEM of a human mandible, including two dental implants, were systematically changed, creating a number of FEMs of the mandible. The effects of the variations were quantified as differences in the principal strain magnitudes modeled by the original FEM (gold standard), prior to the sensitivity analyses, and those generated by the changed FEMs. The material properties that had the biggest impact on the predicted cortical principal strain were the shear moduli (up to 31% in difference from the unchanged state), and the absence of cancellous bone (up to 34%). Alterations to the geometry of the mandibular cross section, such as an increase in corpus dimensions, had the greatest effect on principal strain magnitudes (up to 16%). Changes in the cortical thickness in relation to the width of the corpus section modified strain more than alterations to the corpus depth (14% and 5%, respectively). The relatively small difference (up to 13.5%) between the predicted and measured interimplant distances indicates the accuracy of the FEM. Changes in geometry and orthotropic material properties could induce significant changes in strain patterns. These values must therefore be chosen with care when using finite element techniques for predicting stresses, strains, and displacements. (c) 2006 Wiley Periodicals, Inc.