M Savelieva (1), A Bienczak (1), D Renard (1), N Barbier (1), E Hua (2), S Smeets (1)
(1) Novartis Pharma AG, Basel, Switzerland; (2) China Novartis Institutes for Biomedical Research Co. Ltd. Shanghai, China
Objectives: Integrated approaches to data mining and analytics have the potential to accelerate drug discovery, benefit patients and drive innovation. Early evidence generation allows to optimize labelling, maximize access for patients and enable clinical adoption of a new treatment. Pharmacometric modeling provides a tool for evidence generation throughout all cycles of drug development by e.g., evaluating predictors of response and best treatment options for selected patient subpopulations. Here, we present the modeling and simulation work in support of ligelizumab treatment options optimization for patients with Chronic Spontaneous Urticaria (CSU).
The data collected in ligelizumab Phase 2b study CQGE031C2201 [1] in adult CSU patients who remain symptomatic despite H1-antihistamine treatment was used to:
- Characterize time-course of Urticaria Activity Score over 7 days (UAS7) efficacy response to the background medication (H1-antihistamines) and the factors impacting it
- Characterize time-course of UAS7 efficacy response to ligelizumab treatment and identify predictive factors of response
- Provide model-based predictions of anticipated response in UAS7 change from baseline over time, as well as proportions of well-controlled patients (UAS7 <=6) and Complete Responders (UAS7=0) over time
- Evaluate best treatment options for selected patient sub-populations
Methods:
Longitudinal exposure-response model for weekly Urticaria Activity Score (UAS7) score was developed using Phase 2b adult trial data from study CQGE031C2201 [1]. The overall response to treatment was modeled by first characterizing the UAS7 response over time under “placebo” (H1-antihistamines) and covariates impacting it. Once the final “placebo” model was selected, the respective model parameters were fixed and the UAS7 reduction under ligelizumab was evaluated in the second step, including covariate selection. While performing goodness of fit assessments, special attention was paid to model capacity to capture UAS7 score values close to zero boundary. Simulations using the final model were then performed to evaluate possible dosing adaptations guided by anticipated patients’ response at Week 16 after treatment start.
Results:
The final model selected to describe the exposure-response under ligelizumab treatment for UAS7 time course was an indirect response model with an inhibitory drug effect on UAS7 production. Among covariates tested, race (Asian vs. non-Asian) on Emax and Kout as well as total IgE at baseline on Kout were included in the model. This model was then further used to evaluate an optimal dosing strategy in CSU patients depending on their response to treatment at Week 16. The results of simulation showed that good responders suppress fast and stay suppressed after interval prolongation. Moreover, the dose up-titration scenario for suboptimal responders showed the potential of additional benefit of ligelizumab up-titration for the patients who are not well-controlled (UAS7>6) at Week 16. In particular, doubling the dosing frequency after Week 16 allowed for additional 38% of patients to have their symptoms well-controlled (UAS7<=6) by Week 40 as compared to 16% of patients who remained at the same schedule.
Conclusions: The early use of analytics tools in support of evidence generation plays an increasingly important role. In this work we demonstrated how the model-based assessments of the drug response anticipated for different dosing regimens and populations can be used for tailoring the treatment on example of ligelizumab in CSU patients.
References:
[1] Maurer M, Giménez-Arnau AM, Sussman G, Metz M, Baker DR, Bauer A, et al. Ligelizumab for Chronic Spontaneous Urticaria. N Engl J Med 2019; 381:1321-32.
Reference: PAGE 30 (2022) Abstr 10138 [www.page-meeting.org/?abstract=10138]
Poster: Drug/Disease Modelling - Other Topics