Felix Jost (1), Jakob Zierk (2), Thuy T. T. Le (1), Thomas Raupach (2), Manfred Rauh (2), Meinolf Suttorp (3), Martin Stanulla (4), Markus Metzler (2), Sebastian Sager (1,5)
(1) Institute of Mathematical Optimization, Faculty of Mathematics, Otto-von-Guericke University Magdeburg, Germany (2) Department of Paediatrics and Adolescent Medicine, University Hospital Erlangen, Erlangen, Germany (3) Pediatric Hematology and Oncology, University Hospital “Carl Gustav Carus”, Dresden, Germany (4) Department of Pediatric Hemato-Oncology, Hannover Medical School, Hannover, Germany (5) Health Campus “Immunology, Infectiology and Inflammation (GC-I^3)”, Otto-von-Guericke University, Magdeburg, Germany
Introduction: Acute lymphoblastic leukemia (ALL), characterized by malignant white blood cells (WBCs) and displacement of normal hematopoiesis, is the most common childhood malignancy. For the maintenance therapy (MT), including low-dose treatment of oral 6-mercaptopurine (6MP) and methotrexate (MTX), no international consensus for MT dosing strategies and target levels for WBC and absolute neutrophil count (ANC) suppression exists. Additionally, the intra- and interindividual variability in the pharmacokinetics (PK) and pharmacodynamics (PD) of 6MP and MTX make it challenging to improve the understanding between PD and myelosuppression.
Objectives: We applied and modified a semi-mechanistic population PK/PD model to describe the dynamics of WBCs and ANCs of pediatric ALL patients treated with oral 6MP and MTX during MT [1]. Then, the model was used to simulate different clinically relevant 6MP and MTX dosing regimens and to study the antileukemic effects with its related excessive myelosuppression responsible for delayed, dose-reduced or stopped treatments.
Methods: The data used in this study, 5897 ANCs and 6640 WBC counts, were obtained retrospectively from 116 children (median 4.75 years, 52 (45%) female) who were diagnosed with de novo ALL at university hospitals in Erlangen and Dresden and treated according to the AIEOP-BFM 2000 and 2009 protocols. During MT, patients received oral chemotherapy with daily 6MP (median 40 mg) and once-weekly MTX (median 15 mg) until 2 years after ALL diagnosis. The nonlinear mixed-effects (NLME) modeling was based on the PK/PD model of [1]. During the model development, we replaced the 6MP PK model of [2] with the PK model described by [3] containing a body surface area-dependent clearance to obtain a better response to 6MP dosage. We also tested the influence of weekly MTX administration by either ignoring or considering the administrations and their resulting concentrations through the MTX PK model with a second PD parameter during model fitting. The models were fitted to the data using the first-order conditional estimation algorithm with interaction implemented in NONMEM 7.4. After model fitting, we performed a simulation study in which we compared individual simulated minimal, median, and maximal ANCs resulting from the application of different dosing regimens. The choice of the different doses was based on ALL treatment protocols (AIEOP-BFM 2009 [50 mg/m2], 50% of AIEOP dosis [25 mg/m2], NOPHO-ALL [75 mg/m2], and UKALL [100 mg/m2]).Throughout, we used the fitted models and only varied the chemotherapy dosage.
Results: First, the explicit consideration of MTX within the PK/PD model of [1] only had a minimal/non-significant effect on the model accuracy, so we fixed it to the ratio 2.5:1 between 6MP and MTX and neglected the PK of MTX in the following. Second, our results showed that the use of the PK model of [3] increased the sensitivity of the PD effect and the model accuracy compared to the 6MP PK model of [2]. Third, ANC measurements resulted in higher accuracy than did WBC measurements. The model simulations represented the clinical ANCs quite well in the average and captured trends toward larger or smaller ANC values. However, they did not oscillate as strongly as the measured values. Within the simulation study, a comparison of the four protocols (25, 50, 75, and 100 mg/m^2 BSA 6MP) showed a significant and linear dosage-effect relationship with respect to the total amount of 6MP administered. All (minimal, median and maximal) ANC values decreased linearly, when daily dosing was increased linearly. A tendency for higher oscillations for treatments with pauses and changes in dosage was seen in a comparison of the simulated actual treatment protocol and the constant administrations of the other protocols, which used lower/higher total amounts of 6MP.
Conclusions: We presented a population PK/PD model describing myelosuppression for ALL MT among children who received oral 6MP and MTX. A linear dose-effect relationship superimposed with fluctuations of varying magnitude was shown. Simulations and more mechanistic modeling approaches will allow to improve the understanding of intrinsic and extrinsic influence factors on the aberrant hematopoiesis and chemotherapy-induced myelosuppression of pediatric ALL patients. The monitoring of individual PK profiles and a subsequent analysis of the PK/PD relationship are mandatory next steps for a better dose-effect correlation.
References:
[1] Le, T. T., Jost, F., Raupach, T., Zierk, J., Rauh, M., Suttorp, M., et al. (2018). A mathematical model of white blood cell dynamics during maintenance therapy of childhood acute lymphoblastic leukemia. Mathematical Medicine and Biology doi:10.1093/imammb/dqy017
[2] Jayachandran, D., Rundell, A. E., Hannemann, R. E., Vik, T. A., and Ramkrishna, D. (2014). Optimal chemotherapy for leukemia: a model-based strategy for individualized treatment. PloS One 9, e109623
[3] Hawwa, A. F., Collier, P. S., Millership, J. S., McCarthy, A., Dempsey, S., Cairns, C., et al. (2008). Population pharmacokinetic and pharmacogenetic analysis of 6-mercaptopurine in paediatric patients with acute lymphoblastic leukaemia. British Journal of Clinical Pharmacology 66, 826–837
Reference: PAGE () Abstr 9492 [www.page-meeting.org/?abstract=9492]
Poster: Drug/Disease Modelling - Paediatrics