Abstract
Per FDA guidance, method comparability should be established if an anti-drug antibody (ADA) assay is run by two or more independent laboratories during a study. Genentech, Inc. is evaluating an immunogenicity risk-based comparability approach consisting of both technical and clinical aspects. Technical comparability of the relative sensitivity (RS) is assessed using the Two One-Sided T-tests (TOST) statistical analysis which evaluates if the difference (in absolute value) of the RS means of the two laboratories is less than a pre-specified level of comparability, the practically significant difference (PSD). Clinical comparability is based on the molecule’s immunogenicity risk. A basic and in-depth assessment for low and high-risk molecules are used, respectively. An alternative strategy for molecules with limited incurred sample availability is to be used. In the basic assessment, samples are either unfortified or fortified with surrogate ADA positive control at method appropriate concentrations in a representative biological matrix. Acceptable comparability requires in both methods i) at least 80% of the unfortified samples screen and confirm negative, ii) at least 90% of the low concentration samples screen and confirm positive; and iii) 100% of the high concentration samples screen and confirm positive. The in-depth assessment uses at least 100 incurred samples from 30 or more ADA-positive and ADA-negative patients. The results are evaluated using a 2 by 2-confusion matrix and Cohen’s Kappa score where 1 indicates perfect agreement. Acceptable comparability requires a Cohen’s Kappa score of greater than 0.40. This strategy allows for a robust technical and clinical method comparability assessment. Graphical Abstract