Laure Deyme (1), Florence Gattacceca (1), Laurent Mineur (2), Clémence Toullec (2), Mohamed Gasmi (3), Antonin Schmitt (4), Joseph Ciccolini (1), Marine Gilabert (5) and Dominique Barbolosi (1)
(1) Aix Marseille Univ, INSERM, CNRS, CRCM, SMARTc, F-13005 Marseille, France; (2) Institut Sainte Catherine, Avignon, France ; (3) Service hépato-gastro-entérologie et oncologie digestive, Hôpital Nord, APHM, Marseille, France ; (4) Service Pharmacie, Centre Georges-François Leclerc, Dijon, France; (5) Département d’Oncologie Médical, Institut Paoli Calmettes, Marseille, France
Objectives: Pancreatic adenocarcinoma is one of the most lethal human malignancies and a major health issue. FOLFIRINOX regimen is commonly used in colorectal cancer. Recently, Conroy et al. have shown that FOLFIRINOX is the most efficient regimen in pancreatic cancer. However, this polychemotherapy causes significant and dose-limiting toxic effects leading to empirical dose-reduction, postponement of the forthcoming courses and sometimes treatment discontinuation. The first aim of our study is to describe the PK of each drug included in FOLFIRINOX regimen and to establish dose-concentration-toxicity relationships for haematological and categorical adverse effects. Second, we aim to perform in silico simulation to define the optimal FOLFIRINOX administration protocol (i.e., dosing, scheduling, sequencing) for a maximal benefit to risk ratio.
Methods: As a first step, a multicenter retrospective study was performed. Data from 75 patients with pancreatic adenocarcinoma and 566 courses of FOLFIRINOX (varying doses of each drug according to toxicities) were collected. The dataset contained 596 absolute neutrophil count observations. Patients have often been supplemented with Granulocyte-Colony Stimulating Factor (G-CSF) or pegylated G-CSF (PEG G-CSF) to prevent neutropenia.
The pharmacokinetics of 5-fluorouracil (5FU) was simulated using the population PK model from Terret et al[1]. The PD model for neutropenia was modified from Friberg et al. (2002)[2]. The PKPD model of G-CSF and PEG G-CSF has been added on the Friberg model with effects on proliferation and maturation processes of proliferative cells in bone marrow as described by Pastor et al (2013)[3]. The effect of 5FU chemotherapy has been described with power function (Edrug = slope * Cdrug ^ beta). Five parameters (slope, maximal effect of GCSF on proliferation (Emax1) and on maturation (Emax2), value of free concentration of GSF eliciting 50% of the maximal effect on proliferation (EC501) and on maturation (EC502)) have been estimated using Matlab 2018b. Next, the model has been used to simulate alternative administration protocols of 5-FU and G-CSF targeting lesser use of G-CSF and lesser toxicity.
Results: The pharmacokinetic profile of 5FU was simulated with a two-compartments model and Michaelis-Menten elimination[1] including body surface area as covariate on the maximum rate of elimination (Vmax). The individual PD parameters were estimated. The baseline value of circulating cells was fixed to the observation before the beginning of treatment and PK parameters of drugs (5FU and G-CSF) were fixed. For one patient with nine absolute neutrophil count observations, eight courses of FOLFIRINOX with different doses and six courses of four or five administrations of G-CSF, the following values were obtained: slope = 1.2 L/µg, Emax1=2.87, EC501=0.21 µg/L, Emax2=2.79 and EC502=0.22 µg/L. Several administration protocols were simulated to compare the effect on neutropenia: increasing/decreasing the duration of 5FU perfusion, with/without G-CSF. The PKPD model predicted a better tolerance with a longer duration of 5FU perfusion and in presence of G-CSF.
Conclusion: This preliminary modelling work shows that PKPD modelling can be a useful tool to optimize administration protocol of chemotherapy and G-CSF in FOLFIRINOX regimen. To achieve our goal, effects of oxaliplatin and SN38 will be considered. Furthermore, a Hidden Markov chain Model will be developed to describe major categorical adverse effects (such as peripheral neuropathy and digestive toxicities) in order to find a new FOLFIRINOX regimen that would be more efficient while maintaining all limiting toxicities in an acceptable range.
References:
[1] Terret C, Erdociain E, Guimbaud R, Boisdron-Celle M, McLeod HL, Féty-Deporte R, et al. Dose and
time dependencies of 5-fluorouracil pharmacokinetics. Clin Pharmacol Ther. 2000 Sep;68(3):270–9.
[2] Friberg LE, Henningsson A, Maas H, Nguyen L, Karlsson MO. Model of Chemotherapy-Induced Myelosuppression With Parameter Consistency Across Drugs. J Clin Oncol. 2002 Dec 15;20(24):4713–21.
[3] Pastor ML, Laffont CM, Gladieff L, Schmitt A, Chatelut E, Concordet D. Model-based approach to describe G-CSF effects in carboplatin-treated cancer patients. Pharm Res. 2013 Nov;30(11):2795–807.
Reference: PAGE 28 (2019) Abstr 9064 [www.page-meeting.org/?abstract=9064]
Poster: Drug/Disease Modelling - Oncology