The application of 3D printing technology to pharmaceuticals is expanding, and 3D-printed drug formulations comprising various materials and excipients have been developed using different types of 3D printers. Here, we used a digital light processing-type 3D printer to fabricate poly(ethylene glycol) diacrylate (PEGDA)-based "ghost tablets" that release entrapped drug but do not disintegrate. Three drugs with different aqueous solubilities were incorporated separately into the tablets, and the effects of printer ink composition and printing conditions on tablet formation and drug release were investigated. We also constructed a simple and effective model to predict the drug release profiles of the 3D-printed PEGDA-based tablets based on printer ink compositions and printing conditions. Drug release profiles were constructed by combining data for the amount of drug released at a specified time (15 hr) predicted by a regression algorithm generated by machine learning (multiple linear regression) and the drug release kinetics model generated by a binary classification algorithm (support vector machine). The proposed prediction model is unique and provides information useful for the development of 3D-printed PEGDA-based tablets as future tailored medicines.
Keywords: 3D printing; Controlled release; Digital light processing (DLP); Drug release kinetics; Ghost tablets; Machine learning.
Copyright © 2021 Elsevier B.V. All rights reserved.