Harnessing tumor immunity with chemotherapy: Mathematical modeling for decision-making in combinatorial regimen with immune-oncology drugs.

Abstract

e14095 Background: Combining chemotherapy and immune checkpoint inhibitors (ICI) is challenging due to the near-infinite choice of dosing, scheduling and sequencing between drugs. The aim of this work is to develop a phenomenological model that describes the synergistic effect between cytotoxics and immune check point inhibitors in patients with cancer. Methods: Inspired from literature, we have developed an integrative mathematical model that includes tumor cells, cytotoxic T cells (CTLs) and regulatory T cells (TREGs) plus pharmacokinetics (PK) inputs. Loss in tumor mass is due to combined effect of direct chemotherapy-induced cytotoxicity and CTLs immune response, which is in turn inhibited by the tumor and mitigated by TREGs in the tumor micro-environment. The model describes as well the impact of chemotherapy-induced lymphodepletion on immune tolerance, whereas ICIs protect CTLs against tumor inhibition. Identification of model’s parameters and simulations of various scheduling were performed using Mlxplore software and a Python standalone code. In vitro and in vivo experiments using lung cancer models generate experimental data to adjust model parameters. Results: Complex interplays between cytotoxics and immune cells were best described by a 10-parameters model so as to ensure better identifiability. PK/PD relationships were integrated using compartmental modeling. In silico simulations show how changes in dosing and scheduling impact efficacy endpoints, an observation in line with data from the literature. Ongoing in vitro and in vivo experiments with pemetrexed-cisplatin doublet and anti-PD1 pembrolizumab help optimizing the model’s parameters in a self-learning loop. Conclusions: This work is at the frontier between mathematical modeling and experimental therapeutics with ICIs. In silico modeling and simulations could help narrow down the treatment choices and define optimal combinations prior to running clinical trials. Such model will help identify optimal dosing and scheduling, so as to achieve better synergism and efficacy.