Rapidly and precisely fabricating solid microneedle by integrating vat photopolimerization and machine-learning (VP-ML)

Microneedles (μNs) have emerged as a transformative alternative to conventional hypodermic needles, offering notable benefits such as reduced hazardous waste, lower risk of injury, improved safety, and greater acceptance among individuals with needle phobia. In this study, we introduce an efficient fabrication method combining Vat Photopolymerization and Machine Learning (VP-ML) for μNs production. The method begins with creating a dataset, which is used to develop ML model. This model predicts the printing parameters needed for VP machine based on the desired dimensions. The VP machine works by selectively curing a liquid photopolymer resin layer by layer with a digital mask, enabling precise μNs fabrication. This cutting-edge approach surpasses traditional methods by eliminating the need for molds and complex multi-step processes, enabling faster, customizable fabrication. VP-ML significantly reduces material waste and production costs while yielding sharper needle tips. By integrating Bayesian regularization backpropagation algorithm with ten hidden layers (BR10), the process optimizes printing parameters to enhance production speed and ensure high-quality outcomes. Experimental results demonstrate that VP-ML effectively produces μNs with base diameters greater than 150 μm and heights exceeding 500 μm, achieving a Mean Absolute Percentage Error (MAPE) of less than 10 % among 32 cases demonstrated in this study. Additionally, penetration tests demonstrated that the μNs fabricated using VP-ML were able to penetrate up to 90 % of its total height. These results highlight VP-ML as a green-manufacturing and cost-effective solution for μNs fabrication, while maintaining remarkable improvements in precision, manufacturing efficiency, flexibility, and scalability in the field.