Anne Rodallec1,2, Lee Randy1, Sophie Marolleau1,2, Julien Nicolas3, Sebastien Benzekry1
1COMPutational pharmacology and clinical Oncology Department, Inria Sophia Antipolis – Méditerranée, Cancer Research Center of Marseille, Inserm UMR1068, CNRS UMR7258, Aix Marseille University UM105, 22.
Objectives Paclitaxel (Ptx) is a key chemotherapeutic, but commercial formulations cause toxicity due to emulsifiers. New Ptx formulations aim to improve efficacy and pharmacokinetics (PK). We have developed a subcutaneously (SC) injectable Ptx-polymer (Ptx-PAAm) and propose here a mathematical model to optimize its administration. Methods Pharmacokinetics and pharmacodynamics (PD) studies were done on MCF-7 bearing mice. Estimation of the mixed effects statistical modeling was performed using the stochastic approximation of expectation maximization algorithm implemented in the Monolix® software.The PK model was developed on both intravenous (IV) Ptx and SC Ptx-PAAm (7mg/kg), based on systemic concentration of Ptx for 4 mice per time points (i.e., 0.25, 0.5, 1, 2, 4, 7, 24, 48, and 72 h ). The PD model, focusing on tumor growth was developed on three PD data sets (control, IV Ptx, SC Ptx-PAAm 15mg/kg, n=9 for each group), and validated on an independent group (SC Ptx-PAAm 60mg/kg). Simulations explored multiple treatment schedules including three different total total dose regimens (3960 µg, 6600 µg and 11880 µg, Ptx eq.), administered once, weekly, or daily. In addition, we tested the effect of a loading dose of 1320 µg or 3960 µg (previously determined as the MTD for IV Ptx). Results Based on the objective function, diagnostic plots (VPC, individual fit and prediction distribution), and number of parameters the optimal model included a two-compartment PK structure with linear elimination and no Tlag. Unperturbed tumor growth was best described using reduced Gompertz model and drug resistance in the PD component was needed to properly describe Ptx anti-tumor effect. This resistance was implemented in such a way that the anti-tumor effect of the drug at the first administration could be higher than for later administrations (i.e., E_max1= 6.85 and E_max2=2.51). In these conditions 0 % and 7% outliers for the PK and PD observations, respectively were observed with a maximum of relative standard error of 40% for the fixed effect E_max1. The estimated apparent bioavailability was found to be in the same range as that calculated during non-compartmental PK analysis (i.e., 36% versus 28%, respectively ), as were the AUCs (i.e., 5160 ng/mL/h versus 4631 ng/mL/h and 1350 ng/mL/h versus 1299 ng/mL/h for IV Ptx and SC Ptx-PAAm, respectively). Goodness of fit was assessed for the tumor growth predictions of our test data set (SC Ptx-PAAm 60mg/kg) which was therefore considered as satisfactory. The in vivo results testing the six treatment schedules previously selected by the model, demonstrated excellent agreement with the model predictions. A loading dose (3960 µg) followed by daily administration (417 µg) achieved 60% complete response, outperforming previous internal results (tumor volume reduced to 60 mm³ versus 1350mm³), without additional toxicity (weight loss). Conclusion Our PK/PD model showed SC Ptx-PAAm with optimized regimens significantly increased efficacy over standard schedules.
Bordat A, Boissenot T, Ibrahim N, Ferrere M, Levêque M, Denis S, et al. A Polymer Prodrug Strategy to Switch from Intravenous to Subcutaneous Cancer Therapy for Irritant/Vesicant Drugs. J. Am. Chem. Soc. 2022, 144, 18844 Vaghi C, Rodallec A, Fanciullino R, Ciccolini J, Mochel JP, Mastri M, Poignard C, Ebos JML, Benzekry S. Population modeling of tumor growth curves and the reduced Gompertz model improve prediction of the age of experimental tumors. PLoS Comput Biol. 2020 Feb 25;16(2):e1007178.
Reference: PAGE 33 (2025) Abstr 11647 [www.page-meeting.org/?abstract=11647]
Poster: Drug/Disease Modelling - Oncology