Population PK modelling of camizestrant, a next-generation oral selective estrogen receptor degrader, as monotherapy and in combination with palbociclib in ER+/HER2- advanced breast cancer
Itziar Irurzun-Arana (1), Andy Sykes (1), Srinivas Bachina (2), Tim Brier (3), Justin PO Lindemann (3), Teresa Klinowska (4).
1) Clinical Pharmacology & Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Cambridge, UK. 2) Clinical Pharmacology Data Programming INM, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Waltham, US. 3) Research and Early Development, Oncology R&D, AstraZeneca, Cambridge, UK. 4) Late Development, Oncology R&D, AstraZeneca, Cambridge, UK.
Introduction/Objectives:
Camizestrant (AZD9833) is a next-generation oral selective estrogen receptor (ER) degrader and pure ER antagonist currently under Phase 3 clinical investigation for the treatment of HR+/HER2- metastatic breast cancer.
SERENA-1 (NCT03616587) is an ongoing Phase 1 study evaluating camizestrant in post-menopausal women with ER+/HER2− advanced breast cancer, which has completed its monotherapy (parts A and B) and combination with palbociclib (parts C and D) dose escalation and expansions phases [1] with no maximum tolerated dose identified and demonstrating an encouraging efficacy and dose-dependent safety profile.
The purpose of this analysis was to build a population pharmacokinetic (PK) model to characterize the plasma concentrations of camizestrant and the associated Between-subject variability (BSV) in breast cancer patients following once daily oral administration. We also quantitatively assessed the impact of several intrinsic factors on the PK of camizestrant.
Methods: Longitudinal data from 184 subjects in SERENA-1 (N=108 at monotherapy and N=76 in combination with palbociclib) were used for model building. The data included rich PK information for 5 different doses of camizestrant (25, 75, 150, 300 and 450 mg).
Model development was conducted in NONMEM 7.4 [2] using Stochastic Approximation Expectation-Maximization (SAEM) and Monte Carlo importance sampling (IMP) estimation methods. Different absorption, disposition, variability and covariate models were evaluated. A covariate analysis was performed using different patient characteristics including body weight, age, disease severity or renal/hepatic function. To assess the model performance, graphical model evaluation techniques, such as goodness-of-fit plots (GoFs) and visual predictive checks (VPCs) were used.
Results: Longitudinal plasma concentration-time profiles following camizestrant administration across multiple dose levels were adequately characterized by a population PK model built from rich Phase I study data.
The population PK model consisted of a 2-compartment model with first-order absorption (including 3 additional transit compartments with the same rate constant parameter) and linear elimination. BSV was estimated for key model parameters using an exponential model and an additive error model in log scale described the residual variability. The typical parameter estimates were 2.7 h-1 for the absorption rate and transit rate constants parameters (KA=KTR), 80.5 L/h for the clearance (CL), 1310 L for the apparent volume of distribution in the central compartment (Vc), 718 L for peripheral compartment volume (Vp), 53 L/h for inter-compartment clearance (Q). An effect of dose was included on both CL and Vc in order to account for the supra-proportional increase in exposure observed over the dose range from 25 to 450 mg.
All the fixed effects parameters resulted in a relative standard error <20% indicating precise estimate of parameters and both GoFs and VPCs demonstrated that the model described the data adequately and can be used for simulation purposes. BSV for key model parameters ranged between 33% to 56% coefficient of variation.
The covariate analysis showed body weight (BW) to be significantly correlated with both CL and Vc. Still, the influence of BW was not considered clinically relevant, and hence, no dose adjustment is required for patient based upon weight. The effect of palbociclib on the PK of camizestrant was considered to be not clinically relevant. There was not enough data to assess ethnic differences on the exposure of camizestrant (96% of the population was Caucasian).
Conclusion: A population PK model was successfully developed to describe camizestrant plasma concentrations over time. The effect of palbociclib on the PK of camizestrant was considered to be not clinically relevant.
References:
[1] AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS11-05.
[2] ICON Development Solutions, Ellicott City, MD, USA.