(T1030-01-05) Assessing the Feasibility and Impact of Singlicate-Based Analysis for Clinical ADA Sample Analysis

Purpose: The traditional approach to generating sample results in ligand-binding assays has been to make ‘duplicate’ measurements from two adjacent wells on the same immunoassay plate. In recent years, this paradigm has been challenged, primarily for pharmacokinetic assays, by looking back at previously generated bioanalytical data, with the conclusion that the benefits of duplicate over single well ‘singlicate’ measurement are negligible (Barfield et al. 2020). Feasibility of singlicate analysis for ADA assays has also been evaluated in the method development and the validation space (Stanta et al. 2021; Jiang et al. 2021) where similar conclusions were made. Here we assessed the feasibility and impact of singlicate-based analysis by evaluating sets of clinical ADA sample analysis data in the screening, confirmation and titer tiers of analysis.

Methods: Case Study 1 was a clinical phase I study in healthy volunteers with a bispecific IgG1 monoclonal antibody biotherapeutic modality. The ADA assay format for this study was bridging assay with ECL readout and the assay had a sensitivity of 10 ng/mL based on the polyclonal positive control. 379 ADA samples were analyzed, evaluating the agreement between the result obtained with the sample duplicate versus the results obtained with the 2 individual sample replicates (summarized in Table 1). Case Study 2 was a clinical phase III study in a disease population with a single-chain antibody fragment biotherapeutic modality. The ADA assay format for this study was an Affinity Capture Elution (ACE) assay with ECL readout and the assay had a sensitivity of 100 ng/mL based on the polyclonal positive control. 1122 ADA samples were analyzed, evaluating the agreement between the result obtained with the sample duplicate versus the results obtained with the 2 individual sample replicates (summarized in Table 2).

Results: Case Study 1 In the screening tier, there was agreement between the duplicate and replicate 1 results in 366/379 (96.6%) samples assessed, and between the duplicate and replicate 2 results in 365/379 (96.3%) samples. The signal:noise ratio (SNR) of all samples with a discrepancy was extremely close to the Screening Cut-Point (CP). 33 samples were taken forward to the confirmation tier. Here there was agreement between the duplicate result and replicate 1 result in 32/33 (97.0%) samples, and between the duplicate and replicate 2 result in 30/33 (90.1%) samples. 4 samples were assessed in the titer tier. All samples had the first dilution less than the Titer CP. Example data are presented in Table 3.  Case Study 2 A high level of pre-existing antibodies was present in the study population. In the screening tier, there was agreement between the duplicate and replicate 1 results in 1115/1122 (99.4%) samples assessed, and between the duplicate and replicate 2 results in 1113/1122 (99.2%) samples. Similarly, the SNR of samples with a discrepancy was extremely close to the SCP. 734 samples were taken forward to the confirmation tier where there was agreement between the duplicate result and replicate 1 result in 727/734 (99.0%) samples, and between the duplicate and replicate 2 result in 723/734 (98.5%) samples. The % Inhibition of all samples with a discrepancy was extremely close to the CCP. 560 samples were assessed in the titer tier. There was agreement between the duplicate result and replicate 1 result in 525/560 (93.8%) samples, and between the duplicate and replicate 2 result in 540/560 (98.5%) samples. Reassuringly, the maximum difference in reported titer was only 1 dilution factor. Example data are presented in Table 4.

Conclusion: Our results indicate minimal impact of singlicate analysis on ADA analysis results in the two clinical studies evaluated. In screening and confirmation, only samples that were at the sensitivity of the assay (i.e., close to the SCP or CCP) had any impact on result. In titer, a small number of sample had the result change by only 1 dilution factor. The data provide additional supportive evidence that singlicate-based analysis is feasible in ADA ligand binding assays.

References: Barfield M. et al. European Bioanalysis Forum recommendation on singlicate analysis for ligand binding assays: time for a new mindset. Bioanalysis. 2020; 12(5):273-284.
Stanta J. et al. Comparing singlet and duplicate immunogenicity assay in human plasma for pembrolizumab using Gyrolab®. Bioanalysis. 2021; 13(11):891-900.
Jiang Z. et al. Feasibility of singlicate-based analysis in bridging ADA assay on Meso-Scale Discovery platform: comparison with duplicate analysis. Bioanalysis. 2021; 13(14):1123-1134.