Reinforcement of the Biosimilarity Evaluation in the Phase 3 Study by Incorporating the Phase 1 PK Similarity Evidence - a Bayesian Approach
Eric M. Chi, Amgen Inc.  Hui Wang, .  *Nan Zhang, Amgen 


In order to demonstrate a biosimilar candidate is similar to the reference product, comparative data from analytical, non-clinical, clinical PK and clinical efficacy and safety studies are required. Following the “totality-of-the-evidence” concept suggested by the FDA, strong earlier evidences could entail less stringent requirements for later stage data in the development. However, at this moment, each study in the clinical development (typically a clinical PK study and a clinical efficacy and safety study) has been designed as a stand-alone study (with clinically meaningful margins and rigorous Type 1 error), regardless of the strength of evidence obtained from earlier analytical or non-clinical work. In the assessment of totality of evidence, we focused on a biosimilar clinical development setting with one clinical PK similarity study and one clinical efficacy and safety study and asked how the evidence provided by the clinical PK similarity study could be utilized in making inference on the clinical efficacy and safety study. A Bayesian approach to assess totality of clinical evidence has been developed. This methodology should also enable us to expand and prospectively plan for evidences needed to address residual uncertainty in a stepwise manner for developing biosimilars.