Data analytics on raw material properties to accelerate pharmaceutical drug development.

Area of Science:

• Pharmaceutical Manufacturing
• Data Science in Drug Development
• Materials Science

Background:

• Active pharmaceutical ingredient (API) manufacturability assessment often relies on empirical methods due to limited material availability, posing challenges for process design under tight timelines.
• Efficient utilization of available material data is crucial for accelerating high-quality drug product delivery while minimizing costs and maximizing process capacity.

Purpose of the Study:

• To develop and apply a data-driven methodology for evaluating raw material manufacturability.
• To integrate multivariate analysis and machine learning for informed selection of new materials based on predicted manufacturability.
• To establish a risk assessment tool for early-stage drug product development.

Main Methods:

• A dataset comprising thirty-four APIs and seven excipients was analyzed.
• Eight flow property measurements were collected for each of the forty-one materials.
• Multivariate analysis and machine learning models were employed to cluster materials based on their properties.

Main Results:

• The analysis successfully identified four distinct clusters of materials exhibiting different flow properties.
• The developed models demonstrated the ability to predict manufacturability based on material characteristics.
• The approach facilitates risk assessment by identifying similar surrogate materials for new APIs.

Conclusions:

• The data-driven method enhances the selection of incoming materials for improved manufacturability.
• This approach serves as a valuable risk assessment tool in early product development phases.
• It enables targeted experimentation, optimizing secondary process selection and mitigating key risks.