Daniël Jonker 1, Alina H. Sode 2, Lars-Erik Broksø Kyhl 1, Christian Secchi 3, Pernille Maria Manuel 1, Marcelo Behar 4
1 Ferring Pharmaceuticals A/S (Kastrup, Denmark), 2 PriceWaterhouse Coopers (Hellerup, Denmark), 3 Ferring Pharmaceuticals Inc. (San Diego, USA), 4 PriceWaterhouse Coopers LLP (Sacramento, USA)
Introduction:
Based on 48 years of improving in-vitro fertilisation (IVF) treatment protocols, there is today a deep mechanistic understanding of how treatment with follicle stimulating hormone (FSH) induces multiple follicle growth. Significant progress is being made with individualised treatment protocols, balancing the need for retrieving enough oocytes with the risk of having to cancel a treatment cycle due to hyper response. Despite this, the outcome of ovarian stimulation remains unpredictable and multiple treatment cycles are often required. Further, stimulation with combined FSH and human chorionic gonadotropin (hCG)-driven luteinizing hormone (LH) bioactivity is common practice. In this context there is a need for balancing the amount of LH bioactivity appropriately to ensure optimal follicle growth and development [1].
Objectives:
• To integrate physiological knowledge and clinical trial data into a quantitative systems pharmacology (QSP) model capturing the hormonal regulation of follicle growth and development during ovarian stimulation in agonist and antagonist down-regulation protocols.
• To assess the performance of the QSP model for predicting clinical trial outcomes, with focus on the number of oocytes retrieved and serum progesterone levels.
Methods:
A QSP model was built describing processes spanning the cellular level (such as steroidogenesis and receptor dynamics in theca and granulosa cells), the ovaries (follicle numbers and sizes) and the organism (pharmacokinetics). The model was calibrated using data from two dose-finding trials with recombinant FSH (follitropin delta) and recombinant hCG (CG beta) respectively [2, 1], and with data from a trial comparing agonist and antagonist downregulation protocols head-to-head [3]. The predictive performance of the model was evaluated for a fourth trial conducted with follitropin delta in an Asian population [4]. Distributions of serum anti-müllerian hormone (AMH) level and body weight were simulated based on the trial in-and exclusion criteria and used to derive individual doses of follitropin delta for use in trial simulations. The model is implemented as a system of coupled differential equations on the Jinko platform (Nova In Silico). Approximately 170 compartments, 2.328 species and 7.611 reactions constitute the model framework. Up to 55 follicles/patient are supported.
Results:
The model is initialised 10 days prior to the start of stimulation to establish a pool of follicles with individual sensitivities to gonadotropins, which depend on the AMH level and down-regulation protocol (agonist or antagonist). Follicles are recruited stochastically from this pool and may start growing if adequately exposed to gonadotropins at this time. The fate of each follicle is governed by their dynamically changing sensitivities to gonadotropin, which determine whether the balance shifts towards proliferation or atresia. The key outputs from the model are the number of oocytes retrieved, the duration of stimulation, follicle numbers and sizes, and serum hormone concentrations of progesterone, estradiol, testosterone, androstenedione and inhibin B. The model describes the dose-response relationships of follitropin delta and CG beta for the majority of these trial endpoints accurately and captures agonist and antagonist protocols. The mean number of oocytes in the trial that was not used for calibrating the model was predicted to be 9.2 and is like the observed mean (±SD) of 10.0 ± 6.1. In the same simulation, the median serum progesterone concentration at the end of stimulation was 2.7 nmol/L, which is very close to the observed median (IQR) of 2.4 (1.7; 3.5) nmol/L.
Conclusions:
The QSP model can be used to optimize patient eligibility criteria, choice of down-regulation protocol and treatment regimens involving follitropin delta and CG beta in fertility trials. Owing to its highly mechanistic nature, the model is expected to be useful for exploration of clinical trial outcomes for novel treatment regimens that have not been studied previously.
References:
[1] Fernández Sánchez M, Višnová H, et al. Human Reproduction 2022; 37 (6): 1161-74.
[2] Arce JC, Nyboe Andersen A, et al. Fertility and Sterility 2014; 102 (6); 1633-40.
[3] Lobo R, Sørdal T, et al. Human Reproduction 2024; 39 (7): 1481-94.
[4] Qiao J, Zhang Y, et al. Human Reproduction 2021; 36 (9): 2452-62.
Reference: PAGE 34 (2026) Abstr 12289 [www.page-meeting.org/?abstract=12289]
Poster: Oral: Drug/Disease Modelling - Other Topics