Digital Model‐Driven Cell & Gene Therapy Process Development

The intensification of cell & gene therapy manufacturing processes is crucial for their industrial translation and for addressing large patient populations at a sustainable price, as corroborated by recent communications by the FDA. Achieving this requires innovative process development strategies to reduce production costs while meeting productivity requirements. However, optimization of these processes is a prohibitively time-consuming and resource-intensive task. In this study, we explored a digital model-driven multi-objective optimization framework for cell & gene therapy process development, applied to the expansion of human mesenchymal stem cells in a fixed-bed bioreactor. By integrating a hybrid cell growth model with a cost-of-goods model, we conducted 20,000 in silico experiments to optimize culture time and medium feeding strategies, to maximize productivity while minimizing costs. Our approach yielded a set of processes that balance cost efficiency, yield, and throughput with up to 38% improvement across these objectives compared to an internal benchmark, while reducing process development costs and time by 57% and 81%, respectively. Two of the optimized processes were validated experimentally. This study demonstrates both the feasibility and added value of digital models in optimizing cell & gene therapy manufacturing processes, highlighting them as an effective tool for sustainable, multi-objective process development. By doing so, it addresses critical needs in cell & gene therapy manufacturing and aligns them with recent regulatory encouragement to adopt AI-driven process development approaches.