Peter Bernhardt
Momenta Pharmaceuticals, USA
Title: Fingerprint-like similarity: Making the connection between product characteristics and clinical outcomes
Biography
Biography: Peter Bernhardt
Abstract
Analytical similarity assessments typically focus on comparing the physical, chemical and functional properties of a biosimilars product to the reference product, to understand where the two products are similar and where they are not similar. An attempt is then made to justify why observed differences will or will not impact the clinical outcomes (e.g. safety, including immunogenicity, efficacy, PD or PK). The justification typically requires making a bridge between an observed difference and a clinical outcome via some type of a functional assay. This type of assessment is only as good as the assays being used (e.g. what level of difference would have to be present in order to observe a difference in the functional response). When differences are observed, new or more sensitive assays may be developed, independent of a relationship between the observed
difference and a clinical outcome. We argue that if one starts from a detail understanding of the product, clinical indications,
and safety profile of the reference product and other similar products, then one can a priori determine which physical, chemical
and functional attributes are either known to impact or likely to impact the clinical outcomes. Using this knowledge, attributes
can be identified that may require a detailed understanding of the relationship and interrelationship(s) between the attribute
and the clinical outcomes as well as between attributes and the clinical outcomes (e.g. is there a relationship between high
mannose and sialic acid levels on PK?). With this knowledge first, methods are developed, along with the appropriate controls,
that have the sensitivity required to detect meaningful differences. The assessment of fingerprint like similarity can then be
based on (1) both the full physical and chemical comparison of the product to the RPP and (2) the detailed analysis of the key
properties, with sensitive assays, that allow linkage between key product attributes and clinical outcomes. We will provide an
overview of our view on fingerprint-like similarity and how it can be applied to biosimilars development.