ABSTRACT of the Doctoral Thesis
Surgical treatments of the human mandible are of increasing importance because implants have a high endurance (if implanted correctly), the living comfort with implant borne prostheses is higher than with mucous membrane borne prostheses and the atrophy of the mandibular bone is less. The influence of individual prostheses on the stress state of the mandible is too complex to be estimated with analytical models. In this work methods and algorithms are described to generate Finite Element models based on individual geometry and material properties.
The exact evaluation of the geometry of the bone and the local material properties are of general importance for the quality of the results of such analyses. The generation of Finite Element models based on geometrical items as described in chapter 3 presumes the correct generation of points, curves or surfaces describing the bone or parts of it. With geometrically based models mesh variations can be performed easily and the generation of hexahedral elements is possible, which are favorable for nonlinear analyses considering contact or large displacements. The major drawback with geometrical representations is that geometrical objects such as surfaces or curves are difficult to define correctly for large regions. The fastest way for the Finite Element model generation is to employ the triangularisation used for visualisation purposes as a basis for the mesh generation of the bony surface. In this case the difficulty is to modify the triangularised surface in such a manner that an Advancing Front meshing algorithm can achieve good results. Applying these procedures for the data of a qCT-scan, the Finite Element model of an individual mandible can be generated in a reasonable period of time.
The necessity to assign bone properties that can vary locally and between individuals within a wide range correctly, enforce an automated material detection and property assignment. An evaluation of the literature showed that the local attenuation detected by the qCT corresponds to the local bone density and to the local Young's modulus. A program is described in chapter 4 which assigns material properties to every Finite Element corresponding to the grey values recorded in the qCT-scans.
Muscle directions as detected during the dissectioning of an individual mandible and assigned to the models are compared with data from the literature. Forces generated by the individual masticatory muscles and the assumptions about their physilogical capacity as well as the coordination among them is described in chapter 3.
In chapter 5 several models of human mandibles generated by application of the mentioned routines is described. Models of individual mandibles are generated to evaluate the influence of different surgical treatments such as the variation of the insertion direction of implants or different possibilities of the connection of the implants. Since the load conditions of biomechanical structures are in general complex and vary among individuals it is difficult to create representative loads for a mandible. To give interpretations which are not influenced too much by variations of the individual load conditions, similar treatments for the same structure are analysed and compared. The simulation of the stabilisation of a fracture of a mandible is used to compare the displacements and stress distributions between two comparable treatments. The routines and programs developed for the model generation of the mandible are not restricted to the mandibule, but can also be used for all other bones and for technical structures for the purpose of reverse engineering.