We are developing a multiscale model of the main pathophysiological mechanisms involved in an ischemic stroke. We use this model to carry out virtual experiments in order to better understand the interactions of these mechanisms and their influences on cell damage. These experiments are used to explore the effects of various therapeutic strategies in stroke patients.

We are developing new data-driven tumor growth inhibition (TGI) models for preclinical and clinical data in oncology. These models, based on ordinary differential equations, are more complex than the TGI models found in the litterature. They are aimed at providing a powerful framework for the integration of multiple biomarkers time-course data relative to tumor growth and treatment action.

Data-driven approach consists of building a model strctly relying a given dataset. It is achieved by using regression techniques on available data. It is a classical and powerful modeling approach which may consent to predict the behavior of a system outside the experimental conditions for which the data have been generated. We focus our work on the coupling between powerful statistical regression methods such as mixed-effect techniques and complex biomathematical models of disease progression.

Multiscale approach can be viewed as an alternative modeling method to data-driven approach. It consists of bridging together numerical models to integrate knowledge coming from different scales. These models are mainly theoretical tools but allow for a deep analysis of the complexity of the underlying system. Their investigation through computer simulation can facilitate the design of data-driven models by highlighting the key mechanism to be taken into account.

In oncology, we are working on the development of a multiscale model of tumor growth focusing on the subcellular level (molecular and genetic network), the cell cycle, and the tissue scale by means of partial differential equations coupled to discrete cellular automata framework.