Population based optimal experimental design in cancer diagnosis and chemotherapy: In silico analysis

Interpatient variability is one of the critical issues in the clinical implementation of cancer diagnostic and therapeutic protocols. In this work, model-based population studies are conducted using a tumor–immune model wherein the population is generated in silico by varying the model parameters. This helps us to understand and address the effect of interpatient variability on protocol design. Multi-objective optimization problems are formulated to determine diagnostic and chemotherapeutic protocols for the generated population. The proposed diagnostic protocol directs what to measure and when to measure so that the data is informative to better estimate the parameters influencing the tumor growth. Similarly, a chemotherapeutic protocol is designed for a given population while simultaneously accounting for control of tumor progression and side effects due to doxorubicin. Then the designed chemotherapeutic protocol is applied on the population and the “patients” are classified into two groups (cured and uncured patients) based on the final tumor size. Finally, a classification analysis is done to identify parameter dependent rules that help to predict the success of designed chemotherapeutic protocol. Overall, this kind of in silico analysis will provide some guidelines to choose the most appropriate therapy for a given patient.

► A comprehensive tumor–immune model is considered to perform population based studies.
► Optimization problems are formulated to determine diagnostic and therapeutic protocols for a given population.
► Parameter dependent rules are identified to predict the success of the therapeutic protocol for a given patient.