Machine Learning‐Assisted Fabrication of 3D‐Printed Extracellular Matrix Models

Mimicking the extracellular matrix (ECM) is challenging due to the complex composition, architecture, morphology, and mechanical properties of native tissues, which are key targets in tissue engineering. 3D printing enables the fabrication of ECM models with high spatial resolution exploiting hydrogels retaining essential viscoelastic and water retention properties for cell survival. In this study, a machine learning (ML)-assisted approach was developed to describe the printing behavior of hydrogels, demonstrating the potential to predict printability, despite the limitations imposed by the small available dataset. To generate the hydrogels, gelatin and hyaluronic acid were functionalized with γ-thiobutyrolactone and cysteamine, respectively. Crosslinking was carried out via thiol–ene photochemical reaction with 4-arm-PEG functionalized with norbornene. The resulting formulations were assessed via swelling tests to evaluate their stability, and the most promising candidates were further characterized chemically, morphologically, and rheologically. Cytocompatibility was validated through viability assays using human bone marrow-derived mesenchymal stem cells. High-resolution 3D printing via stereolithography was performed to confirm the printability of the selected hydrogels. Based on these results, a preliminary predictive ML model was developed to estimate and predict hydrogel printability