M.Sc. Laura Herbst, Fraunhofer-Institute for Production Technology IPT, Automation in Life Science
Integrating automation and data-driven modeling is crucial for optimizing cell and gene therapies, particularly in CAR-T cell manufacturing and LNP-encapsulated mRNA production. This study presents two use cases of automation and data-driven approaches: Firstly, a modular, service-oriented AI Agent framework designed to enhance the operational integration of artificial intelligence (AI) in CAR-T cell production. Developed within the EU H2020 AIDPATH project, this framework addresses key requirements such as scalability, regulatory compliance, and process transparency. It facilitates the deployment of AI applications through two applications: a Digital Cell Twin for optimizing cell expansion and a Reactive Online Process Control for monitoring production deviations. Each AI Agent operates autonomously and interfaces with production environments via standardized APIs, ensuring maintainability and adaptability to evolving regulatory standards and technological advancements.
Secondly, a model was developed enabling real-time data exchange and predictive maintenance for mRNA therapeutic production. The integration of a digital twin for managing device states and optimizing production processes highlights the need for standardization in data exchange protocols. Validation in real-world settings underscores the framework’s capability to improve production efficiency and support diverse AI applications. This research underscores the pivotal role of digital twins in ensuring high-quality production and regulatory compliance, positioning them as essential components in the evolution of pharmaceutical manufacturing. Future developments will aim to further automate additional process steps, thereby enhancing the robustness and efficiency of mRNA therapeutic production.
