Owner/Senior Consultant Occams Cooperatie U.A. Twickenham, England, United Kingdom
Disclosure(s):
Jan-Stefan van der Walt: No relevant disclosure to display
Introduction/
Objectives: Immunogenicity, the development of anti-drug antibodies (ADA), is an important characteristic of therapeutic proteins, affecting pharmacokinetics (PK) and pharmacodynamics (PD), including drug efficacy and safety [1,2]. The aim of the analysis was to develop a longitudinal joint PK-ADA model to enable a framework to assess ADA influence.
Methods: Data from 7 clinical trials of avelumab across drug development phases in several tumor types were analyzed using a population approach. Two approaches were investigated: ADA as covariate in a population PK (PPK) model, and Markov models (discrete-time (DTMM) [4] and continuous-time Markov models (CTMM)) [5,6] of ADA status (ADA+ or ADA-). A final PK-ADA model linked the Markov model for ADA status to the PPK model to estimate the bidirectional effects of ADA status on PK and exposure on ADA status simultaneously. Model development and validation was performed using NONMEM (7.4.3) and Perl-speaks-NONMEM (5.3.0).
Results: Data from 1850 patients were split into two subsets stratified by ADA incidence: 1513 (80%) for model development and 337 (20%) for validation. The increase in avelumab clearance (CL) attributable to ADA+ status ranged from 8.5% to 19.9%. The DTMM estimate of the probability of baseline ADA+ status was 2.12%, and to transition from ADA to ADA+ (p01) was 0.7%. The probability to transition from ADA+ to ADA was 25.4%. Baseline ADA status and the relationship between tumor type and p01 were significant covariates. The final CTMM included the effects of baseline ADA status and tumor type on the rate constant for changing from ADA to ADA+ (λ_01) and was 3.8-fold higher when subjects were ADA+ at baseline. Compared to non-small cell lung cancer (most common tumor type), λ_01was 37% lower with urothelial cancer or Merkel cell carcinoma.
The joint PK-ADA model linked the PPK model with the final CTMM (with baseline ADA and tumor type effects on λ_01) by estimating (1) the correlation between avelumab CL and λ_01, (2) effect of probability of baseline ADA+ (p1) on CL (mono-directional joint PK-ADA model), and (3) effects of avelumab exposure on λ_01 and p1 on CL (bi-directional joint PK-ADA model). The relationship between p1 and CL was estimated with maximum CL increase of 11.6% for the mono-directional and 15.0% for the bi-directional model. The maximum decrease in λ_01 was 37%, with 50% of the maximum decrease estimated at an avelumab concentration of 349 μg/mL. Visual predictive checks confirmed that the models were able to predict the validation data with acceptable accuracy.
Conclusions: Baseline ADA status and tumor type were the most important considerations for assessing immunogenicity for avelumab. The bidirectional joint PK-ADA model estimated a lower probability of ADA+ status with an increase in avelumab concentrations.
This study was sponsored by the healthcare business of Merck KGaA, Darmstadt, Germany (CrossRef Funder ID:10.13039/100009945).
Citations: [1] Enrico D, et al. Clin Cancer Res. 2020 Feb 15;26(4):787–92. [2] Collins JM, et al. Hum Vaccines Immunother. 2019;15(4):891–908. [3] Karlsson MO, et al. Clin Pharmacol Ther. 2000 Aug 1;68(2):175–88. [4] Bergstrand M, et al. Clin Pharmacol Ther. 2009;86(1):77–83. [5] Snelder Net al. In: PAGE 18 (abstract 1536). St. Petersburg, Russia; 2009.
