Li Li, Yali Liang, Yan Feng, Amit Roy, Jennifer Sheng
Bristol-Myers Squibb, Princeton, NJ, USA
Assessments of the impact of antidrug antibodies (ADA) on PK, safety, and efficacy are essential to the clinical development of therapeutic proteins. However, these evaluations are complicated by the infrequency of ADA sampling in clinical studies (typically once every 4 or 8 weeks), resulting in unknown ADA status over dosing intervals during which they may impact PK, safety, and efficacy. Moreover, assessments of the effect of ADA are often based on a binary subject level representation of ADA (positive/negative) that do not account for the temporal profile of the ADA or the titer.
Our analysis presents a robust population PK (PPK) model-based approach to assess the impact of ADA on the PK of therapeutic proteins, illustrated with data from the nivolumab clinical development program. The approach refines the previous characterization of the effect of ADA as a time-varying categorical covariate (positive, negative, or missing) [1-3] by evaluating the impact of alternative imputation methods for the unavailable ADA values with a larger dataset, as well as by incorporating ADA titer in the analysis
Objectives:
- Assess the impact of ADA (positive/negative/missing) on nivolumab clearance (CL)
- Assess the sensitivity of alternative imputation methods for the unavailable ADA values on nivolumab CL
- Assess the impact of ADA titer on nivolumab CL
Methods: The present analysis utilized an extensive dataset that included data from 7907 subjects with solid tumors enrolled in 28 clinical studies of nivolumab monotherapy and in combination with ipilimumab. The effect of ADA on nivolumab CL was incorporated as a time-varying covariate in the PPK model, with values of positive, negative, or missing. The sensitivity of the effect to the following alternative imputation methods was assessed: last observation carried forward, impute to have a shortest ADA positive duration, or a longest ADA positive duration. Additionally, log-transformed ADA titer values were tested as a continuous covariate on clearance (CL) using a linear, or hyperbolic (Emax) covariate model. The missing ADA titer was then imputed by last observation carry forward.
Results: The ADA incidence rate was higher for nivolumab in combination with ipilimumab (32%) compared to the nivolumab monotherapy (16%), however the ADA titer was low (< 1:16) in >90% of subjects, with high titer (>1:512 or above) observed in only 0.2% of subjects. Moreover, positive ADA and high titer was present mainly in the first 3-6 months of treatment.
The effect of ADA as a categorical covariate was approximately 20% increase in nivolumab CL (consistent with the previously reported effect [1-3]).
Nivolumab CL was higher in subjects with higher titer. The linear model was chosen to describe the relationship between nivolumab CL and log-transformed ADA titer, based on the lowest BIC. Imputation of missing ADA (positive/negative) or ADA titer duration in modeling led to no significant difference in population parameter estimates. However, the last observation carry forward imputed ADA titer in the population PK model improved individual concentration predictions for subjects with high titer (>1:512 or above).
Conclusions: Nivolumab CL was marginally higher (~20%) when ADA was included as a categorical covariate, and the result was not sensitive to the handling of the unavailable ADA observations (imputed to be missing, positive, or negative). Incorporation of time-varying ADA titer values enabled a more precise assessment of the effect of ADA, and showed that the effect is negligible for ADA titer ≤1:16, and with >20% effect during periods of titer values of ≥1:512.
References:
[1] Bajaj G, et al. Model-Based Population Pharmacokinetic Analysis of Nivolumab in Patients With Solid Tumors, CPT Pharmacometrics Syst Pharmacol. 2017 Jan; 6(1): 58–66
[2] Agrawal S, et al. Evaluation of Immunogenicity of Nivolumab Monotherapy and Its Clinical Relevance in Patients With Metastatic Solid Tumors. J Clin Pharmacol. 2017 Mar;57(3):394-400.
[3] Zhang J, et al. Population Pharmacokinetics of Nivolumab in Combination With Ipilimumab in Patients With Advanced Malignancies. CPT Pharmacometrics Syst Pharmacol. 2019 Dec;8(12):962-970
Reference: PAGE 29 (2021) Abstr 9708 [www.page-meeting.org/?abstract=9708]
Poster: Drug/Disease Modelling - Other Topics