(218c) Data-Driven Process Decisions for Optimization and Robustness: Leveraging Data-Rich Experimentation and Process Modeling

Process development and manufacturing of drug substance material often requires an extensive and detailed experimental strategy to ensure process robustness, successful scale-up, and a suitable control strategy to provide quality active pharmaceutical ingredient (API). However, the burden of generating sufficient data to meet these goals is exacerbated by accelerated project timelines, limited resources, and competing priorities. Consequently, there is a growing need to develop novel experimental approaches and leverage process models to quickly build process understanding and apply the knowledge to make meaningful process improvements.

In this work, we describe methodologies that leveraged automation technologies and data-rich experimentation strategies on multiple synthetic steps in the manufacturing route of an API. By integrating these strategies into the process development of each unit operation, the process knowledge gained from each experiment was maximized in a resource sparring manner. Reaction process fingerprints were collected using process analytical technologies for in situ monitoring and automated sampling technology for off-line chromatographic analysis. The time-series data that was collected from these experiments was used to build mechanistic models and data-driven models to gain critical process insight, optimize operating conditions, and ensure quality drug substance. As needed, empirical relationships were used to scale these laboratory observations to the production environment. Ultimately, the methodology employed in this work enabled a fast and efficient process development to achieve program goals on an accelerated timeline.