Model-Based Drug Development for Oncology Therapeutics

The success rate for drug development in oncology is relatively low compared to other therapeutic areas. This may be due to the many challenges involved in gaining sufficient information from Pharmacokinetic/Pharmacodynamic (PK/PD) quantitative evaluations to effectively guide study design and impact development decisions. This presentation will provide an overview of how PK/PD and Modeling & Simulations (M&S) play an important role in risk-benefit assessments and dose selection for targeted therapies in cancer patients.
Conventionally, anti-tumor activities in early-stage clinical studies are determined from limited information (e.g., the objective response rate), and the dose regimens selected in late-stage development may be based on maximum tolerated dose (MTD), guided safety findings. However, dose limiting toxicity (DLT) may not be reached at the top dose tested in early-stage clinical trials for the targeted therapeutics, particularly for the large-molecule biologics. Therefore, M&S approaches have been implemented to integrate PK/PD information and guide dose selection. First, population PK modeling analyses have been utilized to characterize variability and identify covariates that may affect drug clearance and other parameters. Time-to-event modeling systems have been developed to characterize Kaplan-Meier survival curves, which are then linked to a drug exposure-driven, tumor growth inhibition (TGI) model, and other quantitative/statistical approaches (e.g., Cox regression, Weibull distribution function) to describe exposure-response (E-R) relationships for progression-free-survival (PFS) and overall survival (OS). The unique PK/PD properties of biologics (such as target-mediated drug disposition) and the use of drug exposure as a prognostic vs. predictive factor for E-R relationships should be carefully evaluated. In addition, different types of drug-disease modeling frameworks have been successfully established to predict expected clinical OS outcomes in numerous settings, such as Non-Small Cell Lung Cancer (NSCLC) and other tumor types. This modeling framework focuses on efficacy; it consists of TGI models, with the change in tumor size from baseline and/or the time-to-tumor-growth as the predictor of survival outcomes. This presentation will review different case studies that are based on trial simulations, according to exposure-efficacy/safety relationships, and disease models that support late-stage clinical programs for various cancer types.
In summary, model-based methods integrating PK exposure, tumor dynamics, biomarker, and survival have been implemented to support dose selection for oncology programs. Modeling & simulation may improve product risk-benefit profiles and increase the probability of success for oncology therapeutics.