Hydrogen/Deuterium Exchange Mass Spectrometry (HDX-MS) is a unique, rapidly evolving technique that has been widely adopted for the characterization of the higher-order structure of proteins. Numerous statistical tools are available in the literature that can be used to identify statistically significant differences in the deuterium uptake values. Biopharmaceutical comparability studies, however, require evidence that two protein samples are highly similar and therefore necessitate a different statistical approach. Hageman et al. recently introduced an innovative HDX-MS equivalence testing method utilizing univariate two one-sided tests (TOST). The acceptance limits were established by the randomized resampling of eight replicate measurements of a reference protein. However, this approach can introduce a non-negligible level of variability in the acceptance limits when the same data set is reanalyzed. In the present study, we enhance this method by replacing the randomized resampling process with the systematic enumeration of all possible combinations of the reference data set, thus eliminating the resampling-induced variation. Because this approach incorporates all measurements, including replicate combinations with markedly elevated variability, it leads to higher acceptance limits. Therefore, we evaluated three strategies: robust outlier detection, a percentile-based method, and a partitioning approach to establish more stringent criteria and reduce patient risk. By applying the enhanced methods to data sets of three approved infliximab biosimilars and a partially deglycosylated NIST mAb used as a mock candidate biosimilar, we demonstrated correct classification of equivalent and nonequivalent samples, making the enhanced evaluation strategy well suited for regulatory comparability assessment.
Lénárt et al. (Wed,) studied this question.