Surface-based scaffold design: A mechanobiological approach

Despite recent need-based advances in orthopedic scaffold design, current implants are unsuitable as "total" scaffold replacements. Both mechanical requirements of stiffness/strength and biological stipulations dictating cellular behavior (attachment, differentiation) should be included. The amount of mechanical stimulation in the form of stresses, strains, and energies most suitable toward implant design is presently unknown. Additionally unknown is if whole-bone optimization goals such as uniform and non-uniform driving forces are applicable to a scaffold-bone interface. At the very least, scaffolds ready for implantation should exhibit mechanical distributions (dependent on loading type) on the surface within the typical mechanical usage window. Scaffold micro-architectures can be strategically shifted into that window. The overall goal of this study was to produce micro-architectures tailored to a more uniform mechanical distribution, while maintaining the morphological properties necessary to sustain its mechanical integrity. The mechanical adjustment stimuli investigated were von Mises stress, strain energy density, maximum principle strain, and volumetric strain. Scaffold models of a similar volume fraction were generated of three initial architectures (Rhombitruncated Cuboctahedron, hollow sphere, and trabecular-like bone cube) using high resolution voxel mapping. The resulting voxels were translated into finite element meshes and solved, with a specially written iterative solver created in Fortran90, under confined displacement boundary conditions. The result was verified against a commercial software. Once the mechanical distributions were identified one of two methods was chosen to alter the configuration of material in Cartesian space. The success of the alteration was judged through a diagnostic based on the histogram of mechanical values present on the surface of the micro-architecture. The first method used a compliant approach and, for the case of stress, reinforced locations on the surface with large stresses with extra material (strategically taken from the least stressed portions). The second method used a simulated annealing approach to randomly mutate the initial state in a "temperature" dependent manner. Results indicate that the mechanical distributions of the initial scaffold designs vary significantly. Additionally, the end state of the adjustment demonstrated anisotropy shifts toward the direction of loading. Moreover, the adjustment methods were found to be sensitive both to the mechanical parameter used for adjustment and the portion of the surface adjusted at each increment. In conclusion, scaffolds may be adjusted using a mechanical surface-based objective, as the surface of the scaffold is crucial toward its in vivo acceptance. This technique provides some mathematical specificity toward the whole of computer-aided tissue engineering.