M.Sc. project proposal on learning-based modeling of cancer progression
Understanding the dynamics of brain tumor progression is essential for optimal treatment planning. Cast in a mathematical formulation, it can be viewed as evaluation of a system of partial differential equations, wherein the underlying physiological processes that govern the growth of the tumor, such as diffusion and proliferation of tumor cells, are considered. To personalize the model, i.e. find a relevant set of parameters, with respect to the tumor dynamics of a particular patient, the model is informed from an empirical data, e.g., medical images obtained from different diagnostic modalities, such as magnetic-resonance imaging or positron-emission tomography. The model-observation coupling is typically implemented via statistical inference schemes (e.g. Monte-Carlo sampling).
The general research question we would like to address is whether we can adopt learning-based approaches for the evaluation of the tumor growth. Particularly, we aim to test the methodological advances in parametric inference, relying on deep learning techniques. Besides, we will explore a parallel research line of expressing the whole modeling approach - the tumor simulation and inference tasks - in neural network language. The methods will be validated under models of various biophysical complexity on several retrospective brain tumor studies.
Gained expertise. Computational physiology, deep learning, statistical inference.
Prerequisites. A few self-written lines of code (Python or C++), moderate ability to understand math.