Khalil Ben Hassine (1) *, Sonia Khier (2,3) *, Youssef Daali (4), Tiago Nava (1,5), Henrique Bittencourt (6,7), Maja Krajinovic (6,7), Chakradhara Rao Satyanarayana Uppugunduri (1), Marc Ansari (1,5).
(1) CANSEARCH Research Platform in Pediatric Oncology and Hematology, Department of Pediatrics, Gynecology and Obstetrics, University of Geneva, Geneva, Switzerland; (2) Pharmacokinetic and Modeling Department, School of Pharmacy, Montpellier University, Montpellier, France; (3) Probabilities and Statistics Department, Institut Montpelliérain Alexander Grothendieck (IMAG), UMR 5149, CNRS, Montpellier University, Montpellier, France; (4) Clinical Pharmacology and Toxicology Division, Geneva University Hospitals and University of Geneva, Geneva, Switzerland; (5) Division of Pediatric Oncology and Hematology, Department of Pediatrics, Gynecology and Obstetrics, Geneva University Hospitals and University of Geneva, Geneva, Switzerland; (6) Charles‐Bruneau Cancer Center, Sainte‐Justine University Health Center (SJUHC), Montreal, Quebec, Canada; (7) Department of Pediatrics, Faculty of Medicine, University of Montreal, Montreal, Quebec, Canada; *Equal contribution
Introduction: Busulfan (Bu) remains one of the key components of conditioning regimens in pediatric hematopoietic stem cell transplantation. Typical administration schemes of Bu for myeloablative regimens consist of short intravenous infusions repeated every 6, 12 or 24 hours, for four days. The relationships between Bu exposures (AUC or CSS,avg) and clinical outcomes (e.g. treatment related toxicities) are well described in the literature1,2. Owing to the variability in Bu pharmacokinetics2, therapeutic drug monitoring of Bu became a common practice in most of the pediatric transplantation centers, to accurately target therapeutic cumulative AUC of Bu for each patient.
In addition to the cumulative exposure, per-dose and first-dose Bu exposures are also often associated with toxicities3–6, highlighting the importance of optimizing the accuracy of initial Bu dosing1. To achieve this, Population Pharmacokinetic (PopPK) analyses have attempted to characterize the maturation of Bu clearance. However, patients below 2 years of age require further initial dose improvement. In fact, more than 33% of these patients’ Bu first dose exposures, estimated by current models, are predicted to be outside the EMA therapeutic window (i.e. AUC0-6H of 3.7 – 6.2 mg.h/L). These patients are also at higher risk of treatment related toxicities, such as hepatic sinusoidal obstructive syndrome7, supporting the importance of accurate initial dosing1,2. Physiologically-based pharmacokinetics (PBPK), using the “virtual twin” simulation with pediatric physiological parameters, could be a useful strategy for Bu PK predictions.
Objectives: To assess the performance of a physiologically based pharmacokinetic (PBPK) modelling and simulation approach, for a priori prediction of Bu exposures in pediatric patients, with a special focus on the accuracy of model predictions in patients less than 2 years old.
Methods: A PBPK model was built in Simcyp® software8 based on Bu’s physico-chemical characteristics. Pediatric patients from a prospective observational study (NCT01257854), with available clinical data and measured Bu concentrations, were included in the analysis. A “virtual twin” profile for each patient was developed by individualizing the Simcyp® pediatric population file for the physiological and demographic attributes. Because the ontogeny of Bu metabolizing enzymes (glutathione S-transferases) is not well characterized in the literature, Bu hepatic clearance for each “virtual twin” was calculated based on two previously published PopPK models —formula of Savic et al.9 and formula of Ben Hassine et al.10— according to patients’ individual covariates.
The observed exposures, for the first administration were calculated by non-compartmental analysis (NCA) with PKanalix® software. Then, exposures were simulated (Sim) for each “virtual twin” by the PBPK model. The concordance between observed and simulated AUC0-6H for each patient (i) was then evaluated by the ratio R = AUC(i)Sim/AUC(i)NCA . The performance of simulations was assessed by the mean fold error MFE = AUC(mean, predicted)/AUC(mean, observed) .
Results: Among 209 pediatric patients with available data, 47 were below 2 years of age. For these patients, 47 “virtual twins” were created and AUC0-6H simulated for each Bu clearance formula (Savic et al. or Ben Hassine et al.) based on the administered initial doses.
The ratios R were between 0.75 – 1.25 (equivalent to EMA window) for 85% of the patients when Savic et al. formula was used and 81% when the Ben Hassine et al. formula was used. For ratio limits of 0.85 – 1.15, 60% and 53% of the patients were within the boundaries for Savic et al. and Ben Hassine et al. formula, respectively. For more restrictive ratio limits [0.90 — 1.10], 34% of the patients where within the boundaries for both formulae. The MFE was 0.95 and 1.01 respectively when Savic et al. or Ben Hassine et al. formula was used.
Conclusions: The PBPK-based “virtual twin” simulation strategy has shown to be reliable for predicting the exposure to Bu in patients below 2 years of age even in limited ontogeny data of GSTs. The performance of such an approach shall be compared with those of PopPK predictions. This approach is currently being extended to other pediatric age groups. The strategy could be beneficial in clinics for initial Bu dose recommendations to achieve targeted exposures, by simulating the PK profiles of each patient with different dosing scenarios.
References:
[1] Ben Hassine, K. et al. Total Body Irradiation Forever? Optimising Chemotherapeutic Options for Irradiation-Free Conditioning for Paediatric Acute Lymphoblastic Leukaemia. Front Pediatr 9, 775485 (2021).
[2] Lawson, R., Staatz, C. E., Fraser, C. J. & Hennig, S. Review of the Pharmacokinetics and Pharmacodynamics of Intravenous Busulfan in Paediatric Patients. Clin Pharmacokinet 60, 17–51 (2021).
[3] Ansari, M. et al. Association Between Busulfan Exposure and Outcome in Children Receiving Intravenous Busulfan Before Hematopoietic Stem Cell Transplantation. Therapeutic Drug Monitoring 36, 93–99 (2014).
[4] Benadiba, J. et al. Pharmacokinetics-adapted Busulfan-based myeloablative conditioning before unrelated umbilical cord blood transplantation for myeloid malignancies in children. PLOS ONE 13, e0193862 (2018).
[5] Ansari, M. et al. GSTA1 diplotypes affect busulfan clearance and toxicity in children undergoing allogeneic hematopoietic stem cell transplantation: a multicenter study. Oncotarget 8, 90852–90867 (2017).
[6] Philippe, M. et al. Maximal concentration of intravenous busulfan as a determinant of veno-occlusive disease: a pharmacokinetic-pharmacodynamic analysis in 293 hematopoietic stem cell transplanted children. Bone Marrow Transplantation 54, 448–457 (2019).
[7] Corbacioglu, S., Jabbour, E. J. & Mohty, M. Risk Factors for Development of and Progression of Hepatic Veno-Occlusive Disease/Sinusoidal Obstruction Syndrome. Biol Blood Marrow Transplant 25, 1271–1280 (2019).
[8] Jamei, M. et al. The Simcyp Population Based Simulator: Architecture, Implementation, and Quality Assurance. In Silico Pharmacol 1, 9 (2013).
[9] Savic, R. M. et al. Effect of weight and maturation on busulfan clearance in infants and small children undergoing hematopoietic cell transplantation. Biol. Blood Marrow Transplant. 19, 1608–1614 (2013).
[10]Ben Hassine, K. et al. Precision dosing of IV busulfan in pediatric hematopoietic stem cell transplantation: Results from a multicenter population pharmacokinetic study. CPT Pharmacometrics Syst Pharmacol (2021).doi:10.1002/psp4.12683
Reference: PAGE 30 (2022) Abstr 10207 [www.page-meeting.org/?abstract=10207]
Poster: Drug/Disease Modelling - Paediatrics