Semi-mechanistic pharmacokinetic-pharmacodynamic modeling of tumor size dynamics in advanced breast cancer patients treated with single-agent amcenestrant
Marc Cerou (1), Hoai-Thu Thai (1), Laure Deyme (2), Sylvaine Cartot-Cotton (3), Christine Veyrat-Follet (1)
(1) Translational Disease Modelling Oncology, Sanofi, 91380 Chilly-Mazarin, France (2) Modeling and Simulation, Sanofi, 34184 Montpellier, France (3) Pharmacokinetics Dynamics and Metabolism Sanofi, 91380 Chilly-Mazarin, France
Introduction: Selective estrogen receptor degraders (SERDs) provide an important therapeutic option for hormone receptor positive breast cancer [1]. SERDs act by fully antagonizing and degrading the estrogen receptor (ER), resulting in inhibition of the ER signaling pathway. Amcenestrant is an oral SERD in development for the treatment of ER+, HER2- breast cancer [1]. Tumor growth inhibition (TGI) modeling aims to describe the dynamics of tumor size evolution, anti-tumor drug effect as well as resistance to treatment. Several approaches have been developed to characterize drug resistance [2] from empirical models, e.g., Claret model [3], to more mechanistic models with intra-tumor heterogeneity [4].
Objectives: The objectives of this study were (i) to characterize the effect of amcenestrant as single agent on tumor size dynamics in phase 1 study patients with ER+ and HER2- advanced breast cancer (ii) to evaluate the amcenestrant dose-response relationship and (iii) to identify baseline covariates impacting tumor response
Methods: Data from 2 phase 1 clinical studies, AMEERA-1 (NCT03284957 - Parts A+ B) [5] and AMEERA-2 (NCT03816839) [6] of single agent amcenestrant in non-Japanese and Japanese patients with ER+ and HER2- advanced breast cancer were included in this analysis. Tumor size data (sum of target lesions diameters selected based on RECIST criteria v1.1) were analyzed using different structural TGI models. Treatment exposure over time was introduced using concentrations predicted by a population pharmacokinetic model characterized by two-compartment distribution, linear elimination and four transit compartments to account for absorption delay. Once the best structural TGI model was selected, the relationship between baseline covariates and model parameters were evaluated using univariate analysis followed by the automatic covariate selection procedure (COSSAC) [7] implemented in Monolix 2020R1. Trial simulations (100 trials of 1000 patients with the same baseline characteristics to those in AMEERA-1 and 2) were then performed to evaluate the benefit of different dosing regimens (100, 200, 400, 600 mg QD) regarding target lesion response rate (TLRR). Model parameters were estimated using the SAEM algorithm implemented in Monolix2020R1, and simulations were performed using SimulX2020R1 and R version 3.6.1.
Results: A total of 305 tumor size measurements in 75 patients from AMEERA-1 and AMEERA-2 studies were considered. The time course of longitudinal data of tumor size of target lesions was best described by a semi-mechanistic TGI model with 2 compartments accounting for the dynamics of sensitive & resistant cells. Amcenestrant acts by inhibiting the tumor proliferation of sensitive cells, and its effects were best characterized by an Imax model. A delay on the initiation of drug effect was observed and considered in the model using two effect site compartments. The covariate analysis identified 5 baseline covariates which have impacts on tumor size dynamics: patients without liver metastasis and lower number of organs with metastasis (NMET<3) tend to have lower tumor size at baseline, as compared to patients with liver metastasis or higher number of organs with metastasis (≥3). Patients tend to have faster tumor regrowth if they have low albumin (35 g/L) or presence of lymph node metastasis or prior CDK4/6 inhibitor therapy, as compared to patients with median albumin (41 g/L) or absence of lymph node metastasis or without prior CDK4-6 inhibitor therapy.
Simulations using the final model indicate that 400 mg QD provides a slightly higher TLRR vs. 200 mg QD (+2.3 % at 12 months) and higher TLRR vs. 100 mg QD (+4.8% at 12 months) and a plateau at 600 mg QD is observed with no additional benefit compared to 400 mg QD (+1.2% at 12 months).
Conclusion: The present TGI model characterized the tumor size dynamics well on both the individual and population levels. Based on the simulations, a limited positive dose-response relationship on tumor size of target lesions was predicted in this advanced breast cancer population. However, progression from non-target or new lesions should be considered in future work to allow objective response rate prediction according to RECIST 1.1 criteria and to refine the dose-response assessment.
References:
- Chen et al. Latest generation estrogen receptor degraders for the treatment of hormone receptor positive breast cancer. Expert Opin Investig Drugs. 2021: 25 (1-15).
- Sun X and Hu B. Mathematical modeling and computational prediction of cancer drug resistance. Brief Bioinform. 27 nov 2018;19(6):1382‑99.
- Claret L et al. Model-Based Prediction of Phase III Overall Survival in Colorectal Cancer on the Basis of Phase II Tumor Dynamics. J Clin Oncol. 1 sept 2009;27(25):4103‑8.
- Yin A. A Review of Mathematical Models for Tumor Dynamics and Treatment Resistance Evolution of Solid Tumors. Syst Pharmacol. :18.
- Linden HM et al. A Phase 1/2 study of amcenestrant (SAR439859), an oral selective estrogen receptor (ER) degrader (SERD), as monotherapy and in combination with other anti-cancer therapies in postmenopausal women with ER-positive (ER+)/human epidermal growth factor receptor 2-negative (HER2–) metastatic breast cancer (mBC): AMEERA-1. San Antonio Breast Cancer Symposium 2020.
- Kotani H et al. AMEERA-2: Phase 1 study of oral SERD amcenestrant (SAR439859) in Japanese women with ER+/HER2- advanced breast cancer. The Japanese Society of Medical Oncology Annual Meeting 2022.
- Ayral G et al. A novel method based on unbiased correlations tests for covariate selection in nonlinear mixed effects models: The COSSAC approach. CPT Pharmacometrics Syst Pharmacol. 2021 Apr;10(4):318-329.
