Machine learning enabled integrated formulation and process design framework for a pharmaceutical 3D printing platform
Machine learning enabled integrated formulation and process design framework for a pharmaceutical 3D printing platform
复制标题
机器学习为制药 3D 打印平台提供集成配方和工艺设计框架
DOI：
10.1002/aic.17990
复制
发表时间：
2022
期刊：
影响因子：
3.7
通讯作者：
Reklaitis, Gintaras V.
中科院分区：
文献类型：
--
作者：
Sundarkumar, Varun;Nagy, Zoltan K.;Reklaitis, Gintaras V.
制药制造业需要快速发展，以吸收下一波颠覆性工业创新——工业 4.0。这涉及结合人工智能和 3D 打印 (3DP) 等技术来实现药品生产过程的自动化和个性化。本研究旨在为制药 3DP 平台构建配方和工艺设计 (FPD) 框架，该框架建议可以获得一致的液滴打印的操作（配方和工艺）条件。本研究中使用的平台是基于位移的按需滴落 3D 打印机，通过将药物制剂以液滴形式沉积在基材上来制造剂量。 FPD 框架由两部分组成：第一部分涉及构建机器学习模型来模拟正向问题（预测给定操作条件下的打印机操作），第二部分寻求解决并通过实验验证逆向问题（预测可产生所需打印机操作的操作条件）。
The pharmaceutical manufacturing sector needs to rapidly evolve to absorb the next wave of disruptive industrial innovations—Industry 4.0. This involves incorporating technologies like artificial intelligence and 3D printing (3DP) to automate and personalize the drug production processes. This study aims to build a formulation and process design (FPD) framework for a pharmaceutical 3DP platform that recommends operating (formulation and process) conditions at which consistent drop printing can be obtained. The platform used in this study is a displacement‐based drop‐on‐demand 3D printer that manufactures dosages by additively depositing the drug formulation as droplets on a substrate. The FPD framework is built in two parts: the first part involves building a machine learning model to simulate the forward problem—predicting printer operation for given operating conditions and the second part seeks to solve and experimentally validate the inverse problem—predicting operating conditions that can yield desired printer operation.