(T-107) Comprehensive Human Quantitative Systems Pharmacology Models for Immuno-Oncology: Enhancing Decision-Making on Target Indications and Combination Strategies for New Drug Development

Kota Toshimoto, Ph.D – Scientist, Astellas Pharma Inc; Hiroyuki Sayama, Ph.D – Team Lead, Astellas Pharma Inc.; Georgi Kapitanov, Ph.D – Director, Applied BioMath/Certara; Soumya Rao, Ph.D – Principal Scientist, Applied BioMath/Certara; Fei Hua, Ph.D – VP, Applied BioMath/Certara; Kalyanasundaram Subramanian, Ph.D – CEO, Differentia Biotech; Karyn Sutton, Ph.D – Associate Director, Applied BioMath/Brystol Myers Squibb; Masayo Oishi, Ph.D – Head, Astellas Pharma Inc.

Objectives: In oncology drug development, given the mechanism of action (MoA) of drug candidates, it is critical to select appropriate target cancer types and develop combination strategies with potential standard of care (SoC). Quantitative systems pharmacology (QSP) models for immune-oncology (IO) incorporating mechanisms of cancer-immunity cycle have been recognized as powerful tools to assist in this process, and various IO-QSP models have been published. However, comparing anticipated clinical responses across different cancers can be challenging due to publicly available models offering varying structures across cancer types. This study aims to establish platform IO-QSP models with a unified structure for gastric and pancreatic cancers, facilitating direct comparison of the effect of drug-candidate combinations with SoCs.

Methods: A published IO-QSP model for immune checkpoint inhibitors in breast cancer served as the base model [1]. To expand the base model to gastric and pancreatic cancers, cancer-specific parameters such as initial tumor diameter, tumor growth rate and immune cell proportions were collected from literature and integrated into the model. Virtual clinical trials were conducted to simulate anti-PD-1 monotherapy and anti-PD-1 + anti-CTLA-4 combination therapy. The predicted responses were compared to clinical data for model calibration and validation. Subsequently, three SoC treatments were incorporated into the expanded models: FOLFOX and ramucirumab plus paclitaxel for gastric cancer, and FOLFIRINOX for pancreatic cancer. To streamline the SoC model structure, only two MoAs were assumed: tumor growth inhibition and apoptosis induction, with corresponding scaling factors for each MoA (GF and DF, respectively) incorporated into the model. The values for GF and DF were optimized for each SoC treatment to capture the clinical responses of it. Virtual clinical trials for combination therapy of each SoC with anti-PD-1 were then conducted and compared to clinical observations for final model validation.

Results and

Conclusion: A total of 12 cancer-specific parameters for gastric and pancreatic cancers were collected and integrated into the model, successfully reproducing the clinical responses of anti-PD-1 and/or combination therapy with anti-PD-1 and anti-CTLA-4. Optimal combinations of GF and DF were identified for each SoC, effectively reproducing clinical responses. Finally, the models were validated by confirming the reproducibility of clinical responses for the combination of each SoC with anti-PD-1 treatment. Overall, the established IO-QSP models captured clinical responses of SoCs and checkpoint inhibitors in both gastric and pancreatic cancers well, thus having the potential to contribute to decision-making regarding target indications and combination strategies for new drug development.

Citations: [1] Wang H, et al. Conducting a Virtual Clinical Trial in HER2-Negative Breast Cancer Using a Quantitative Systems Pharmacology Model With an Epigenetic Modulator and Immune Checkpoint Inhibitors. Front Bioeng Biotechnol. 2020;8:141.