Regression-guided computational design of auxetic scaffolds for soft tissue applications

The mechanical performance of tissue-engineered scaffolds plays a critical role in their effectiveness for regenerative medicine. While auxetic metamaterials offer tunable mechanical behavior ideal for soft tissues, their design typically relies on inefficient, iterative trial-and-error processes. To address this limitation, this study presents an integrated computational framework for the inverse design of auxetic scaffolds. By combining Finite Element Method (FEM) simulations with regression-based models, we developed accurate predictive models capable of mapping microstructural parameters directly to macroscopic mechanical responses. This data-driven approach allowed for the rigorous optimization of four distinct auxetic architectures to replicate the complex, non-linear anisotropic properties of human skin, achieving strong agreement with literature targets. A primary contribution of this work is the development of a user-friendly software tool that integrates this pipeline. The tool allows users to input target mechanical properties and automatically generates optimized, fabrication-ready designs (including custom MEW G-code), effectively bridging the gap between theoretical metamaterial optimization and practical clinical application. This methodology supports robust, patient-specific scaffold development, significantly advancing the capabilities of soft tissue engineering. • Novel inverse design framework automates the generation of patient-specific auxetic scaffolds. • Data-driven regression replaces iterative trial-and-error, predicting mechanics with high accuracy. • Computational tool bridges the gap between design and manufacturing, generating fabrication-ready files. • Four auxetic microarchitectures are optimized to replicate the non-linear anisotropy of human skin. • The integrated workflow enables rapid, target-driven customization for soft tissue engineering.