Health problems related to bones (e.g., trauma, osteoporosis, tumors) incur considerable societal costs. Each year in the U.S. alone, there are an estimated 6.5 million bone defects recorded. To this end, cost- effective and sustainable solutions for synthetic bone implants catering to patient-specific needs is vitally important. Current state-of-the-art synthetic bone implants based on metamaterials like trusses and plates have their own set of challenges. Natural bone has an  incredibly complex microstructure which spans multiple length scales and is heterogeneous in porosity, anisotropy, and mechanical stiffness within the same specimen and across different patients. Therefore, the ideal implant design must locally match these properties of the natural bone (specific to the patient and the anatomical site in question) in a spatially-variant fashion for improving long-term compatibility and preventing bone atrophy. 

Our approach bypasses this challenge and applies an inverse model trained on spinodoid metamaterials to the in-silico generation of artificial trabecular bone. The inverse model accurately predicts topologies that match the target relative density and anisotropic elastic stiffness as well as bear resemblance to the natural bone topologies, and achieves this, remarkably, without prior information about bone during the learning stage.

Ref: S. Kumar, S. Tan, L. Zheng, D. M. Kochmann, Inverse-designed spinodoid metamaterials, npj Comput. Mater., 6 (2020), 73.