Preview

PDF, English Download (11MB) | Terms of use

Abstract

The first results of the German pediatric precision oncology program INdividualized Therapy FOr Relapsed Malignancies in Childhood (INFORM) showed the significance of high evidence levels for successfully matched targeted therapy based solely on molecular diagnostics. Yet, only a small number of patients (8%, 42/519) (1) actually present with a high evidence target, highlighting an unmet need to improve drug response predictions and clinical treatment recommendations. Therefore, the aim of this thesis is to integrate pharmacodynamic (PD) parameters from Drug Sensitivity Profiling (DSP) with pharmacokinetic (PK) parameters, and improve drug response prediction in high risk pediatric patients. To achieve this aim, a literature review was conducted, and nine PK parameters focused on the pediatric population were collected for the drugs from the DSP drug library in the INFORM study. In addition, a database of primary patient tumor (PPT) samples (n=68) and a database of positive control cell (PCC) line models (n=7) were generated. The PCC models habor a specific molecular alteration (e.g., BRAF V600E, NTRK fusion) with a clinically proven drug- target relationship. Among the 68 PPT samples, five samples (PPT subgroup I) harbored a very high priorty (INFORM priorty score 1) alteration with a clinically proven drug-target relationship. Both the PPT samples and PCC models underwent DSP using a library of 79 clinically relevant oncology drugs. Hit selection was based on dose-response curves-derived PD parameters and PD-PK integrated parameters. These parameters were evaluated for their predictive value in the PCC models and the PPT subgroup I samples. Subsequently, the parameter with the best predictive value was investigated in the PPT samples without a defined drug-target relationship. A PK database of 74 drugs and nine PK parameters for each drug focusing on the pediatric population was successfully created and published for the scientific community. When investigating the predictive power of PD parameters, the drug sensitivity score (DSS) z-score showed the best predictive power in identifying the matching drug in the PPT subgroup I samples based on the molecular background. However, the DSS z-score could not capture the patient's clinical history. Conversely, the integrated PD-PK parameter, the DSS Cmax z- score, could effectively capture the patient's clinical history in the PPT subgroup I samples. In the PPT samples without a defined drug target match and no clinical treatment history, the DSS Cmax z-score provided additional insights for 77% (n=53/68) of the patient samples that were not detected by NGS molecular analysis. In summary, a previously unavailable and comprehensive pediatric PD database was generated and published to serve the scientific community. The PK parameter Cmax was identified and successfully integrated with the DSS, introducing a novel DSP metric for drug response prediction. The groundwork established by testing and describing the DSS Cmax z- score in this thesis serves as a foundation for further investigation in larger datasets with clinical outcomes. This could refine the prediction of drug response for pediatric high-risk patients and improve their treatment selection without relying on time-consuming and costly techniques.