(M-070) Population pharmacokinetic and exposure-response analysis of high-dose methotrexate-associated mucositis in infant patients with acute lymphoblastic leukemia

Zachary Taylor – Cincinnati Children's Hospital Medical Center; Erin Breese – Cincinnati Children's Hospital Medical Center; Tamara Miller – Emory University; Austin Brown – Baylor College of Medicine; Eric Schafer – Texas Children's Hospital; Brooke Bernhardt – St. Jude Children's Research Hospital; Sharon Castellino – Children's Healthcare of Atlanta; Maureen O'Brien – University of Colorado Anschutz Medical Campus; Laura Ramsey – Children's Mercy Kansas CIty

Objectives: High-dose methotrexate (HDMTX) is a cornerstone of treatment for pediatric acute lymphoblastic leukemia (ALL), yet is associated with marked pharmacokinetic (PK) variability and toxicity. Infants receive empiric dose reductions due to delayed drug elimination from developmental immaturity, but the effectiveness of these strategies in mitigating toxicity remains unclear. This study aimed to develop a population PK model and investigate exposure–response relationships for HDMTX-associated mucositis in infants with ALL.

Methods: PK and toxicity data were retrospectively collected from infant, pediatric, and adolescent-young adult patients treated with HDMTX (3.5–5 g/m² over 24 hours) at three pediatric institutions. Infants were defined as ≤1.5 years of age. A 3-compartment population PK model (ADVAN11 TRANS4) was developed in NONMEM, incorporating body surface area, baseline serum creatinine, time-varying creatinine change from baseline, and categorical age as covariates on clearance. Model performance was evaluated using prediction-corrected visual predictive checks and nonparametric bootstrap analysis. Post hoc Bayesian concentration estimates were generated for each patient. Mucositis events were graded per CTCAE v5.0 following manual chart review. Logistic regression was used as an inferential tool to evaluate exposure–response relationships and estimate HDMTX concentrations associated with 50% and 75% probabilities of mucositis.

Results: This ALL cohort comprises 2125 administrations from 572 non-infant patients and 150 administrations (6.6%) from 49 infant patients (7.9%). The final model identified age-related reductions in clearance: infants < 6 months, 6–12 months, and 12–15 months had 25.1%, 18.2%, and 9.1% lower clearance, respectively, compared to non-infants. Infants had a higher incidence of any-grade mucositis (OR: 2.32; 95% CI: 1.64–3.28; p < 0.0001). The 20-hour methotrexate concentration best described mucositis risk (BIC: 3001.5). In infants, the 50% and 75% probability thresholds were 40.8 µM (95% CI: 29.9–49.0) and 96.9 µM (95% CI: 87.9–107.9), compared to 98.9 µM and 154.4 µM in non-infants. Exploratory modeling also identified the AUC24-36 (BIC: 2994.9) and peripheral (V3) concentrations from 36 hours onward (BIC: 3000.2) as exposure metrics associated with mucositis.

Conclusions: This population pharmacokinetic and inferential exposure–response analysis demonstrates that infant patients, despite current empiric dose reductions, exhibit slower HDMTX elimination and greater susceptibility to toxicity. The 20-hour plasma concentration, AUC24-36, and peripheral compartment concentrations starting at hour 36 offer mechanistic insight into delayed distribution kinetics and their association with mucositis. These findings support the reevaluation of current infant dosing protocols and warrant prospective studies to further characterize HDMTX distribution kinetics across a range of patient ages and body sizes.

Citations: 1. Beechinor RJ, Thompson PA, Hwang MF, et al. The Population Pharmacokinetics of High-Dose Methotrexate in Infants with Acute Lymphoblastic Leukemia Highlight the Need for Bedside Individualized Dose Adjustment: A Report from the Children's Oncology Group. Clin Pharmacokinet. 2019;58(7):899-910. doi:10.1007/s40262-018-00734-0

Keywords: Pediatrics, Oncology, Exposure-Response