DDEL-01. The role of key pharmacodynamic and pharmacokinetic parameters in drug response prediction of pediatric tumors in the precision oncology study INFORM

Abstract

Abstract INTRODUCTION: The INFORM (INdividualized Therapy FOr Relapsed Malignancies in Childhood) study is a European pediatric precision oncology program. Using state of the art molecular assays, INFORM aims for identification of targetable genetic alterations, matching drugs and clinical trials. High evidence targets were associated with doubling of progression free survival when patients received a matching drug. However, the fraction of tumors with high evidence drug targets remains low requiring functional layers of information such as drug sensitivity profiling. The aim of this project is to identify and investigate the role of key pharmacodynamic and pharmacokinetic parameters to improve the predictivity of ex vivo drug response of pediatric tumors. METHODS: Positive control cell lines harboring specific mutations (n=7) and primary tumors (n=121) from INFORM, including 10% ependymomas, 7% high grade gliomas, 5% neuroblastomas and 4% medulloblastomas, were profiled ex vivo using a library of n=76 clinically relevant oncology drugs in a 384 well plate format. Metabolic activity was measured after 72h of treatment. Quality control (QC) was done using the robust z-factor, correlation of replicates and mean negative control. Hit selection was based on maximum percentage inhibition, normalized AUC metric (DSSasym) and maximum serum concentration (Cmax) of the drug. Clinical follow-up was collected using a questionnaire. RESULTS: A linear mixed model revealed the DSSasym to be the strongest pharmacodynamic parameter in drug prediction in cell lines. Drug screens of n=105 INFORM cases passed QC. Application of the filtering parameters resulted in prediction of n=1-16 drugs/case (min-max). A data base of published pediatric pharmacokinetic parameters of the drug library was generated. Analysis of predictive parameters and clinical follow-up of clinical samples is ongoing. CONCLUSION: Including pharmacodynamic as well as clinical pharmacokinetic parameters is paramount to identify potentially clinically active compounds from ex vivo drug screen data. Further algorithm development is warranted.