M. van Eijk (1), H. Yu (2), T.P.C. Dorlo (1), J.H. Beijnen (1,3), A.D.R. Huitema (1,4,5)
(1) Department of Pharmacy & Pharmacology, Netherlands Cancer Institute - Antoni van Leeuwenhoek, Amsterdam. (2) Pharmacometrics, Clinical Development & Analytics, Global Drug Development, Novartis, Basel. (3) Utrecht Institute of Pharmaceutical Sciences, Utrecht University, Utrecht. (4) Department of Pharmacology, Princess Máxima Center for Pediatric Oncology, Utrecht. (5) Department of Clinical Pharmacy, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands
Introduction: The oral administration of the widely used anticancer agent paclitaxel has several benefits compared to the conventional intravenous (IV) administration [1]. Other than the ease of administration, prolonged plasma exposure above threshold concentrations may be more easily attained. For single agent IV paclitaxel it has been demonstrated that an increased time during which paclitaxel plasma concentrations exceed 0.05 µmol/L (42.7 ng/mL) is a good predictive marker of both myelosuppression and longer time to disease progression [2]. The use of oral paclitaxel in a low-dose metronomic (LDM) schedule may even further facilitate the achievement of longer time above threshold concentrations. Formulation and co-administration with ritonavir are key in achieving clinically relevant plasma concentrations after the oral administration of paclitaxel. Moreover, it has been hypothesized that LDM treatment with paclitaxel may lead to upregulation of the matricellular protein thrombospondin-1 (TSP-1), which has been proposed as a potential mediator of the effects of LDM chemotherapy [3].
Objectives:
- Development of a pharmacokinetic (PK) model, in which we quantify the different effects of various formulations and co-administration with ritonavir on paclitaxel PK parameters.
- Investigation of the pharmacodynamic (PD) relationship between with TSP-1 and LDM chemotherapy with oral paclitaxel.
Methods: PK data was available for 58 patients, treated with three different oral paclitaxel formulations: a drinking solution, a freeze-dried solid dispersion capsule formulation and a spray dried solid dispersion tablet formulation. PD data was available for 36 patients who were treated with oral paclitaxel in a LDM schedule. A previously developed model was used to describe oral ritonavir pharmacokinetics [4]. Empirical Bayesian parameter estimates from this model were used in the development of the oral paclitaxel model to describe the effect of ritonavir on hepatic paclitaxel metabolism using a well-stirred liver model. Different absorption models were explored to describe paclitaxel absorption. Comparison between the different formulations was based on the parameter estimates of relative gut bioavailability and absorption rate. Population PK/PD analysis was performed using NONMEM version 7.3.0. [5]. Pharmacokinetic simulations based on population parameter estimates were performed in R (version 4.0.3) using the developed oral paclitaxel model and a previously established model for IV paclitaxel [6, 7].
Results: For paclitaxel a semi-physiological model consisting of gut, liver, central and peripheral compartments was developed. Paclitaxel gut absorption was best described using a Weibull function with different parameter estimates for the solid dispersion formulations and drinking solutions. The model incorporated the inhibition of hepatic paclitaxel metabolism proportional to ritonavir plasma concentrations with an estimated inhibition constant of 560 ng/mL (95% CI: 296 – 1066). The apparent uninhibited intrinsic clearance was estimated as 881 L/h (95% CI: 632 – 1250). The relative gut bioavailability of the tablet formulation was similar to that of the drinking solution. The capsule formulation had a 46% lower gut bioavailability (95% CI: 28% – 60%) compared to the paclitaxel drinking solution. Simulations demonstrated that the oral paclitaxel tablet formulation achieves a longer cumulative time above 0.05 µmol/L per 3-week cycle compared to weekly IV administration when applied in a low dose metronomic schedule (177 h. vs. 31 h.). The PK/PD relationship between paclitaxel and TSP-1 was modelled using a turnover model in which paclitaxel plasma concentrations increase the TSP-1 input rate following an Emax relationship with an estimated EC50 of 230 ng/mL (RSE 45%).
Conclusions: We successfully developed a PK/PD model that accurately describes the complex PK of oral paclitaxel co-administered with ritonavir. Herein, we quantified the fundamental effects of formulation and co-administration with ritonavir on paclitaxel PK and quantified a PD relationship with TSP-1.
References:
[1] Rowinsky E. N Engl J Med. 1995;332(15):1004–14.
[2] Joerger et al. Clin Cancer Res 2007;13 (21)
[3] Bocci et al. Proc Natl Acad Sci U S A. 2003;100(22):12917–22.
[4] Yu et al. J. Clin. Pharmacol. 2020, 60 (3) 340–350
[5] Beal SL et al. 1989-2011. NONMEM Users Guides. Icon Development Solutions, Ellicott City, Maryland, USA.
[6] Development Core Team R. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2008
[7] Joerger et al. Clin Cancer Res. 2006;12:2150-7.
Reference: PAGE 29 (2021) Abstr 9837 [www.page-meeting.org/?abstract=9837]
Poster: Drug/Disease Modelling - Oncology