Two-step global sensitivity analysis of a non-local integro-differential model for Cancer-on-Chip experiments

Area of Science:

• Computational biology
• Mathematical modeling
• Immunology

Background:

• Cancer-on-chip models simulate tumor-drug interactions and immune responses.
• Understanding parameter influence is crucial for model accuracy and experimental design.

Purpose of the Study:

• To perform a global sensitivity analysis on a non-local integro-differential model of cancer-on-chip experiments.
• To identify key model parameters affecting immune cell dynamics and spatial distribution.
• To assess the model's reliability and suggest improvements for accuracy and efficiency.

Main Methods:

• A two-step global sensitivity analysis was employed, starting with the screening Morris method.
• The extended Fourier Amplitude Sensitivity Test (eFAST) was used to quantify parameter importance.
• 13 model parameters and 11 target outputs monitoring immune cell behavior were analyzed.

Main Results:

• The Morris method identified a core set of 6 influential parameters out of 13.
• eFAST quantified the significant impact of chemical field and cell-substrate adhesion parameters.
• Non-linear effects and parameter interactions were investigated.

Conclusions:

• The model is feasible within the explored parameter space.
• Chemical signaling and cell-substrate adhesion are critical factors for model accuracy.
• Model simplification and improved experimental data can enhance predictive reliability.