Steps Toward Model-Informed Precision Dosing of PEG-Asparaginase in Children and Young Adults
Jordan Brooks (1), Jasmine Hughes (1), Vanessa Burns (1), Keri Zabokrtsky (2), Van Huynh (2), Sonia Morales (2), Ron Keizer (1)
(1) InsightRx, (2) Children’s Hospital of Orange County
Introduction: Polyethylene glycol-conjugated asparaginase (PEG-ASNase) is integral in the treatment for acute lymphoblastic leukemia (ALL) in children and young adults as it improves the overall survival of these patients. However, treatment-limiting toxicities (including hypersensitivity, pancreatitis, and liver dysfunction) are prevalent in children and young adults receiving PEG-ASNase, and subtherapeutic levels of PEG-ASNase are associated with inferior disease-free survival [1,2]. An additional concern to health system payers is PEG-ASNase notable cost, which in the United States can be up to 20,000 -30,000 USD per dose. Würthwein et al. published a population pharmacokinetic (PopPK) model of PEG-ASNase, in which they describe the notably high inter- and intrapatient variability present in the treatment population [3]. Additionally, Kloos et al. have proposed a TDM framework for children receiving PEG-ASNase to improve treatment success, which is currently accessible in real-world care as the measurement of ASNase activity levels (AAL) is already standard practice in assessing for inactivation of ASNase by anti-ASNase antibodies [4].
Objectives:
In this study, we aim to:
- Prospectively validate and refine the PopPK model developed by Würthwein et al. in a novel PEG-ASNase data set.
- Using simulations, evaluate the ability of model-informed precision dosing (MIPD) to achieve therapeutic target attainment.
Methods: Patient data were collected from children and young adults receiving PEG-ASNase for treatment of ALL collected as a part of a clinical trial at Children’s Hospital of Orange County (CHOC). Data cleaning and visualizations were created in R and nonlinear mixed effects modeling and simulations were completed in NONMEM utilizing PsN, including the proseval tool for prospective evaluation, assessed in terms of mean percentage error (MPE) and normalized root-mean square error (nRMSE) with uncertainty calculated via bootstrap. Simulations of several dosing strategies were evaluated: traditional body surface area (BSA) based dosing, MIPD a priori dosing, percentage-based dose adjustments as suggested by the algorithm developed by Kloos et al., and MIPD a priori and a posteriori dosing. Dose adjustments were based on simulated AAL including residual variability. For MIPD, initial dose selection targeted an AAL of 0.3 IU/mL based on the population model and patient covariates. For subsequent doses, if the measured concentration was outside of the acceptable range (0.1 – 0.5 IU/mL), MAP Bayesian estimates of pharmacokinetic parameters for each patient were used to re-adjust the dose to attain the target of 0.3 IU/mL.
Results: A total of 143 patients with 667 AAL were included in the final data set. The median age of the patients was 7.5 y (1.1 - 23.9), 38% were female biologic sex, 11 patients had a BMI > 30 kg/m2 (obese), 15 had a BMI 25.5-29.9 (overweight), and 25 (3.7%) of AAL were below the limit of quantification. The Würthwein model adequately described the data and refitting the model found comparable structural parameter estimates and proportional and additive error terms; however, higher interindividual variability on volume of distribution was estimated. The refit model had a significantly improved objective function (dOFV -7.1785, p<0.01) and a relatively improved visual predictive check. Prospective evaluation of the refit model found a median nRMSE of 33.5 % (IQR: 32.7 – 34.3) and a median MPE of -24.9 % (IQR: -46.3 – -15.7). Simulations suggest the use of MIPD with dose adjustments improves target attainment. For fixed FDA guideline dosing without dose adjustment, simulations indicate up to 2.2% will experience subtherapeutic levels <0.1 IU/mL and 68.8% will experience levels >0.5 IU/mL; whereas, MIPD suggest 0.4% will be <0.1 IU/mL while preventing any levels > 0.5 IU/mL at steady state. Additionally, simulations suggest an average lower dose across the length of therapy if adopting AAL-informed dose adjustments, translating to lower treatment costs without compromising therapeutic target attainment.
Conclusions: Adoption of MIPD in the routine clinical use of PEG-ASNase in children and young adults being treated for ALL has real potential to improve therapeutic level attainment while lowering treatment costs. Furthermore, this study externally validates an existing PopPK model, highlighting that the tools for implementing MIPD in clinical care are available.
References:
[1] Bender C, Maese L, Carter-Febres M, Verma A. Clinical Utility of Pegaspargase in Children, Adolescents and Young Adult Patients with Acute Lymphoblastic Leukemia: A Review. Blood Lymphat Cancer. 2021 Apr 19;11:25-40. doi: 10.2147/BLCTT.S245210. PMID: 33907490; PMCID: PMC8064615.
[2] Van Thu Huynh, Carol Hwang Lin, Sonia Morales, Anurag K. Agrawal, Christine L Phillips, Keith J. August, Erin M. Guest, Beth Apsel Winger, Catherine Aftandilian, Janel R. Long-Boyle, Kirsten Kasper, Keri Zabokrtsky; Asparaginase-Associated Toxicities and Hypersensitivity Reactions in Pediatric and Adolescent Down Syndrome Patients with Acute Lymphoblastic Leukemia or Lymphoma. Blood 2023; 142 (Supplement 1): 5809. doi: https://doi.org/10.1182/blood-2023-190734
[3] Würthwein G, Lanvers-Kaminsky C, Siebel C, Gerß J, Möricke A, Zimmermann M, Stary J, Smisek P, Schrappe M, Rizzari C, Zucchetti M, Hempel G, Wicha SG, Boos J; AIEOP-BFM ALL 2009 Asparaginase Working Party. Population Pharmacokinetics of PEGylated Asparaginase in Children with Acute Lymphoblastic Leukemia: Treatment Phase Dependency and Predictivity in Case of Missing Data. Eur J Drug Metab Pharmacokinet. 2021 Mar;46(2):289-300. doi: 10.1007/s13318-021-00670-8. Epub 2021 Feb 17. PMID: 33595793; PMCID: PMC7935823.
[4] Kloos RQH, Pieters R, Jumelet FMV, de Groot-Kruseman HA, van den Bos C, van der Sluis IM. Individualized Asparaginase Dosing in Childhood Acute Lymphoblastic Leukemia. J Clin Oncol. 2020;38(7):715-724. doi:10.1200/JCO.19.02292
