Development of statistical methods for analytical similarity assessment

Area of Science:

• Biopharmaceutical analytics
• Regulatory science
• Statistical methodology

Background:

• Evaluating analytical similarity is crucial for biosimilar product approval.
• The US-licensed reference product serves as the benchmark for comparison.
• Current methods require a robust framework for assessing quality attributes.

Purpose of the Study:

• To present a tiered statistical approach for analytical similarity assessment.
• To define criticality of quality attributes (QAs) based on risk.
• To detail statistical methods for three tiers of QA evaluation.

Main Methods:

• A tiered approach starting with QA criticality determination.
• Risk ranking of QAs based on impact on product quality and clinical outcomes.
• Application of three tiers of statistical approaches based on QA criticality.

Main Results:

• A framework for classifying QAs based on their criticality.
• Identification of three tiers of statistical methods with varying stringency.
• Detailed discussion of the Tier 1 equivalence test for analytical similarity.

Conclusions:

• The proposed tiered approach provides a structured method for biosimilar analytical similarity assessment.
• Criticality determination of QAs is key to applying appropriate statistical methods.
• Statistical challenges in equivalence testing for biosimilars require further consideration.