An efficient statistical approach to design 3D-printed metamaterials for mimicking mechanical properties of soft biological tissues

Using 3D printed, patient-specific medical phantoms has become increasingly popular for use in biomedical applications including medical device testing, medical education, and surgical planning, etc. To overcome the inherent differences in mechanical properties between biological tissues and printable polymers, metamaterials are being introduced to mimic the mechanical response of the biological tissues. However, the existing trial-and-error approaches for finding the geometric parameters of the metamaterial result in time-consuming trials, which cannot meet the urgent needs for medical applications. We addressed this issue by proposing an optimization-based statistical approach with an easy-to-evaluate surrogate model to guide the design process and reduce the design time. In this paper, several validation tests were reported, including a biomedical application of mimicking the mechanical response of human articular cartilage. The proposed approach achieves excellent accuracy both visually and quantitatively. In addition, we provide an analysis of mimicking different stress-strain curves using different metamaterials. This data-driven approach demonstrates efficacy and flexibility in building the surrogate model even when no obvious physical trends can be extracted. With the proposed statistical approach, we can efficiently design the metamaterial and 3D-print mechanically accurate phantoms for sophisticated engineering applications. © 2018