Comparison of geometry-based and CT voxel-based finite element modelling and experimental validation.

The objectives of the paper presented here is the comparison of a geometry-based and voxel-based finite element (FE) method preprocessor of the human femur. The models were experimentally validated by strain gauge measurements (principal stress). The correlation coefficients (r) between the three methods (geometry-based FEM, voxel-based FEM, strain gauge measurements) were found to be in the range 0.91-0.94 (r2: 0.84-0.88). The relationships between the samples are highly significant (P = 0.001), where the strain gauge results are the independent variables. These results suggest that the validity with respect to the principal stress of a voxel-based modelling is similar to the validity of geometry-based modelling. In summary, therefore, we conclude that voxel-based meshing allows a straightforward interfacing with computerized tomography (CT) scans and might contribute to a clinically applicable FEM technology.