Abstract
Bioprinting is an emerging technology for the construction of complex three-dimensional (3D) constructs used in various biomedical applications. One of the challenges in this field is the delicate manipulation of material properties and various disparate printing parameters to create structures with high fidelity. Understanding the effects of certain parameters and identifying optimal parameters for creating highly accurate structures are therefore a worthwhile subject to investigate. The objective of this study is to investigate high-impact print parameters on the printing printability and develop a preliminary machine learning model to optimize printing parameters. The results of this study will lead to an exploration of machine learning applications in bioprinting and to an improved understanding between 3D printing parameters and structural printability. Reported results include the effects of rheological property, nozzle gauge, nozzle temperature, path height, and ink composition on the printability of Pluronic F127. The developed Support Vector Machine model generated a process map to assist the selection of optimal printing parameters to yield high quality prints with high probability (>75%). Future work with more generalized machine learning models in bioprinting is also discussed in this article. The finding of this study provides a simple tool to improve printability of extrusion-based bioprinting with minimum experimentations.