In Silico Strategies for Robust Process Development in Advanced Therapies: Poster Presented at PDA Week 2026

Area of Science:

• Biotechnology and Biopharmaceutical Manufacturing
• Process Analytical Technology (PAT)
• Computational Biology

Background:

• Process development for cell and gene therapies, such as CAR-T, faces challenges due to inherent patient biological variability.
• Robust process development is crucial for enhancing product quality and reducing lifecycle costs.

Purpose of the Study:

• To illustrate an integrated approach for analyzing complex cell therapy processes using in silico methods alongside experimental data.
• To quantify the contribution of patient-specific versus process-controlled variability to overall outcome variance.

Main Methods:

• Design of Experiments (DOE) to define experimental space.
• Mechanistic modeling to predict critical quality attribute (CQA) outcomes.
• Machine learning (regression, random forest) to train models.
• Monte Carlo simulation to quantify variability and assess risk.

Main Results:

• An integrated framework combining DOE, mechanistic modeling, machine learning, and Monte Carlo simulation was applied to a simulated case study.
• The approach successfully quantified the impact of patient variability on CQAs.
• Simulations demonstrated the ability to ensure CQAs remain within acceptable limits despite patient input variations.

Conclusions:

• The integrated approach offers more robust process development than DOE alone for personalized cell therapies.
• This framework aids in guiding process optimization and risk assessment from early development to commercial production.
• In silico analysis alongside experiments is key to managing complexity in autologous cell therapy manufacturing.