In the process development of pharmaceutical active pharmaceutical ingredients, there is a need to establish robust manufacturing methods to ensure the stable supply of high-quality pharmaceuticals at low cost. This presentation will introduce two case studies of the use of statistical optimization methods using JMP and JMP Pro at Kyorin Pharmaceutical's CMC Research Institute.
The first example is an analysis of an optical resolution process using the response surface method. Experiments were conducted with the aim of identifying important factors and proving robustness, and statistically, the cooling temperature was identified as a critical process parameter that affects yield and quality (optical purity).
The second example is the rapid optimization of reaction conditions using Bayesian optimization. Under the urgent need for scaling up, we successfully optimized four factors—temperature, solvent volume, and others—simultaneously, finding the target conditions in just 15 experiments.
In all cases, the derived optimal conditions showed good reproducibility in scale-up experiments from 50g to 500g. Through this presentation, we will introduce how data-driven decision-making utilizing experimental design and Bayesian optimization contributes to the efficiency of pharmaceutical process development.
