Tassadit Belabbas1, Ida Neldemo2, Lukas Kovar1, Mehdi Lahmar1, Angela Märten1, Alejandro Pérez-Pitarch3, Celine Sarr2, Ulrike Schmid1, Reinhard Sailer1, David Busse1
1Boehringer Ingelheim Pharma GmbH Co. KG., 2Pharmetheus AB, 3Boehringer Ingelheim Pharma GmbH Co. KG., Ingelheim, Germany; at time of project involvement, now: Regeneron Pharmaceuticals, Inc., Tarrytown, NY
Introduction: Brigimadlin (BI 907828) is a murine double minute 2 (MDM2)-tumor suppressor protein p53 (TP53) (MDM2-p53) antagonist developed for the treatment of locally advanced and metastatic solid tumors. Nonclinical data suggests that inhibition of the MDM2-p53 protein interaction results in cell cycle arrest and/or apoptosis in tumors harbouring wild-type TP53 but not in mutant TP53. The main dose-limiting toxicities reported were myelosuppressive events, including thrombocytopenia. Objectives: We aimed to characterize the relationship between brigimadlin pharmacokinetics (PK) and platelet count via modelling and simulation by integrating relevant data across the program, with the objective of supporting the dose optimization of brigimadlin for future patient populations. Methods: The population modeling analysis included data from a Phase I (NCT03449381) and a Phase II study (NCT05512377). A total of 4982 quantifiable PK observations and 5747 platelet count observations from 283 patients (152 males, 131 females), were integrated for model development. Individual concentration-time profiles, reflecting the actual dose given and potential dose delay, were predicted from a population PK model, and a semi-physiological model for myelosuppression [1] was used as base model. Modeling analysis of the platelet count was performed with non-linear mixed effect modeling (NONMEM® version 7.5.1) using first-order conditional estimation method with interaction (FOCE–I) throughout the model development [2]. Perl-speaks-NONMEM version 5.3.1 (PsN) was used for automation and post processing [3]. Data management and further processing of NONMEM outputs was performed using R version 4.3.1. The model evaluation was based on goodness-of-fit plots and visual predictive checks (VPCs). Using the final model, platelet time course over two cycles was simulated (n=1000) for two oral doses (30 mg vs 45 mg Q3W). The proportions of thrombocytopenia grade 4 and dose delay (platelet count < 100 x 10^9 cells/L) were compared between the two dosing regimens. Thrombocytopenia grades were defined according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5) [4]. Results: The final model consisted of a proliferating compartment linking the PK model to a safety model, three transit compartments, and a compartment for the circulating platelets. The model is characterized by two drug effects. The first is a sigmoidal maximum effect model with maximum effect on the proliferating platelets, fixed to 1. The second is a linear drug effect describing a removal from the third transit compartment. Furthermore, addition of a time effect related to the platelet feedback parameter allowed a better prediction of the platelet time course including the nadir. The typical values of the baseline platelet count, and mean maturation time (MTT) were 204 x 10^9 cells/L (RSE: 2.5%) and 173 h (RSE:12.2%), respectively, while brigimadlin concentration at half maximum effect (EC50) was 2670 nmol/L (RSE: 9.8%). Based on the simulations results, 13.8% of the patients were predicted to develop thrombocytopenia grade 4 following 45 mg Q3W compared to 7.3% for the 30 mg Q3W dosing regimen. Conclusions: By integrating all available data from relevant Phase I and Phase II trials, an exposure-safety analysis was performed, assessing the relationship between brigimadlin PK and the occurrence of severe thrombocytopenia. The developed semi-mechanistic model was successfully applied to simulate the platelet time course and potential dose delay. The target dose of 45 mg Q3W is expected to result in an increased risk of grade 4 thrombocytopenia compared to 30 mg. For the purpose of dose optimization, these results need to be interpreted in context of other relevant safety endpoints as well as anti-tumor activity for the two assessed dose levels.
[1] Friberg LE, Henningsson A, Maas H, Nguyen L, Karlsson MO. Model of chemotherapy-induced myelosuppression with parameter consistency across drugs. J Clin Oncol. 2002;20: 4713–4721. doi:10.1200/JCO.2002.02.140 [2] Beal SL, Sheiner LB, Boeckmann AJ, Bauer RJ. NONMEM 7.4 Users Guides. (1989–2019). https://nonmem.iconplc.com/nonmem744. Gaithersburg, MD: ICON plc; 2019.y across drugs. J Clin Oncol. 2002;20: 4713–4721. doi:10.1200/JCO.2002.02.140 [3] M. Bergstrand, A.C. Hooker, J.E. Wallin, M.O. Karlsson. Prediction-corrected visual predictive checks for diagnosing nonlinear mixed-effects models. AAPS J, 2011;13: 143-151. doi: 10.1208/s12248-011-9255-z [4] Common Terminology Criteria for Adverse Events (CTCAE) Version 5. https://ctep.cancer.gov/protocoldevelopment/electronic_applications/docs/ctcae_v5_quick_reference_5x7.pdf. US Department of Health and Human Services, National Institutes of Health, National Cancer Institute. 2017
Reference: PAGE 33 (2025) Abstr 11339 [www.page-meeting.org/?abstract=11339]
Poster: Drug/Disease Modelling - Oncology