Image-based anatomical reconstruction and pharmaco-mediated bone remodeling model applied to a femur with...

Patient-specific finite element (FE) models are constructed in a way that boundary conditions, material and geometric FE models respect the unique characteristics of the patients, offering a more reliable and realistic structural analysis. This study presents an image-based construction process of a patient-specific FE model applied to a femur with a subtrochanteric fracture. Available procedures are systematized and described in greater depth aiming to investigate the difficulties and possibilities in the development of this type of model. A set of image processing techniques (segmentation, labeling, morphological operations and registering) were performed for volumetric reconstruction and registration of femoral fragments by assessing computed tomography (CT) images in a semi supervised procedure. The CT numbers feed the FE mesh with a non-homogeneous distribution of elasticity modulus. A FE analysis and a bone remodeling simulation with pharmacological stimulus based on the dynamics of bone cells populations were also developed for this study. The FE analysis revealed that the principal stresses provided results close to those obtained using a generic homogeneous model, with a difference of about 2%. In contrast, principal strains presented significant differences, up to 12 and 21 % for the maximum tensile and compressive principal strain values, respectively. The critical values of tensile principal strains were located in the superior part of the femoral neck, a typical region of femoral fractures, and maximum principal strains 2 and 3 (compressive) were concentrated in the diaphyseal region, rather close to the actual fracture. These results corroborate previous works indicating that the strain-based failure criterion is more successful in capturing the bone fracture sites when compared to stress-based criteria. The attempt to reconstruct the material state before the failure as implemented here constitutes an innovative way to analyze and reconstruct bone fractures. In the bone remodeling simulation, the inclusion of an antiresorptive agent stimulus produced cortical bone thickening, an increase in the BMD biomarker, and an overall increase in the bone ash density, just as expected in an osteoporosis therapeutic treatment. In conclusion, this study highlights the advantages of subject-specific FE models compared to generic ones in the study of abnormal clinical cases, evaluation of fracture risk, and optimization of surgical intervention. On the other hand, the high level of human intervention and expertise involved in patient-specific FE analyses hinders the incorporation of this technique into clinical practice.