Abstract:
|
Multiple companion diagnostics could be available for testing a biomarker, for example PD-L1 testing assays. Different assays could lead to different treatment decisions, it will be interest in understanding concordance between two assays. Besides the comparator companion diagnostic(CCD), there is an opportunity to develop a follow-on companion diagnostic(FCD) in use with the same therapeutic drug. This study is to establish the concordance between different assays at balanced cutoffs using the method proposed by Li (Li,2016). The null hypotheses will test the agreement between FCD and CCD is inferior to the intra-agreement of two CCD replicates by a proposed non-inferior margin for both positive(PPA) and negative percentage agreements(NPA). Sample size will be calculated based on the estimated agreement between FCD and CCD and within two replicates of CCD, non-inferiority margin, power and type I error rate etc. We applied this method to understand the range of sample size required to power a concordance study to establish concordance and harmonization between multiple assays approved with different therapeutic products. This method can be generalized to other diagnostic assays.
|