An Integrated Population Pharmacokinetic Model Versus Individual Models of Depatuxizumab Mafodotin, an Anti‐EGFR Antibody Drug Conjugate, in Patients With Solid Tumors Likely to Overexpress EGFR

• Published in: Journal of clinical pharmacology 2019-09, Vol.59 (9), p.1225-1235
• Main Authors: Mittapalli, Rajendar K., Stodtmann, Sven, Friedel, Anna, Menon, Rajeev M., Bain, Earle, Mensing, Sven, Xiong, Hao
• Format: Article
• Language: English
• Subjects: ABT‐414; Computer applications; Conjugates; depatux‐m; Epidermal growth factor; Epidermal growth factor receptors; Epitopes; Glioblastoma; Glioblastoma multiforme; Pharmacokinetics; population PK; Radiation; Solid tumors; Temozolomide; Tumors
• ISSN: 0091-2700; 1552-4604
• DOI: 10.1002/jcph.1418

Depatuxizumab mafodotin (depatux‐m) is an antibody‐drug conjugate (ADC) designed for the treatment of tumors expressing epidermal growth factor receptor (EGFR), consisting of a veneered “humanized” recombinant IgG1κ antibody that has binding properties specific to a unique epitope of human EGFR with noncleavable maleimido‐caproyl linkers each attached to a potent antimitotic cytotoxin, monomethyl auristatin F. We aimed to describe the development and comparison of 2 population pharmacokinetic modeling approaches. Data from 2 phase 1 studies enrolling patients with glioblastoma multiforme or advanced solid tumors were included in the analysis. Patients in these studies received doses of depatux‐m ranging from 0.5 to 4.0 mg/kg as monotherapy, in combination with temozolomide, or radiation plus temozolomide depending on the study and/or arm. First, an integrated ADC model to simultaneously describe the concentration‐time data for ADC, total antibody, and cys‐mafodotin was built using a 2‐compartment model for ADC for each drug‐to‐antibody ratio. Then, 3 individual models were developed for ADC, total antibody, and cys‐mafodotin separately using 2‐compartment models for ADC and total antibody and a 1‐compartment model for cys‐mafodotin. Visual predictive checks suggested accurate model fitting across a range of concentrations. The analysis showed that both an integrated complex ADC model and the individual models that have shorter computational time would result in similar outcomes.