Planning Split-Apheresis Designs for Demonstrating Comparability of Cellular and Gene Therapy Products

Area of Science:

• Biostatistics
• Biotechnology
• Regulatory Science

Background:

• FDA draft guidance addresses statistical considerations for cell and gene therapy product comparability.
• The split-apheresis design is one experimental approach discussed.
• Current recommendations suggest a paired data analysis for this design.

Purpose of the Study:

• To evaluate the statistical power of paired analysis for split-apheresis designs.
• To investigate methods for improving power in comparability studies.
• To provide statistical approaches for incorporating historical data.

Main Methods:

• Demonstrated statistical underpowering of paired analysis for specific quality attributes with small sample sizes (3-5 donors).
• Assessed the impact of donor variability on statistical power.
• Developed and presented statistical methods for integrating historical data from pre-change lots.

Main Results:

• Paired analysis is underpowered for certain quality attributes with limited donors.
• Statistical power increases when a substantial portion of variability is donor-attributable.
• Incorporating historical data significantly enhances power for these attributes.

Conclusions:

• Paired analysis alone may be insufficient for split-apheresis comparability studies with small sample sizes.
• Utilizing historical data through appropriate statistical methods is crucial for robust comparability assessments.
• The study provides practical statistical solutions for enhancing power in cell and gene therapy comparability studies.