Evolutionary Predictions In Colorectal Cancer (EPICC)

We aim to make prognosticating cancer like forecasting the weather. Weather forecasting combines detailed measurement of the current atmospheric state with a mechanistic understanding of atmospheric evolution that can be ‘played forward’ using a mathematical model to give accurate predictions. In oncology, we have the capability to make detailed measurements of the current state of a cancer, but critically lack a mechanistic understanding about how tumours will evolve over time. The shortfall in knowledge presents a major hurdle to accurate prognostication and is the focus of our proposal. We will perform a uniquely high-resolution molecular analysis of human colorectal cancers, and via mathematical modelling of these data, derive an unprecedented quantitative understanding of the ‘evolutionary laws’ that underpin colorectal carcinogenesis. We will evaluate the prognostic value of these laws, and then use them to construct and test models that mechanistically forecast disease evolution. The proposal builds upon our previous work demonstrating the predictability of cancer genomic alterations (Williams et al., Nature Genetics, 2016) as a direct consequence of physical constraints on tumour evolution (Sottoriva et al., Nature Genetics, 2015). This research will represent a major step towards the replacement of correlation-based prognostication with a new paradigm of accurate mechanistic forecasting.