One hallmark of cancer is its rapid and relentless growth. But that growth is uneven: some areas of a tumour grow quickly and others grow slowly. This variability in behaviour also exists between individuals: some people have tumours that respond well to therapy and can be cured; others, sadly, do not. In fact, tumours that look superficially very similar can have very different responses to treatment. This grant focuses on understanding tumour heterogeneity by considering how tumours differ at the microscopic level. In particular, different parts of a tumour show large regional differences in blood circulation, and that different patients have very large differences in oxygenation. Tumours whose blood vessels are inefficient at providing oxygen are well-known to respond very poorly to many therapies. This project aims to understand why. I will develop mathematical models that predict tumour oxygenation (as measured experimentally) by identifying the genetic factors that differ between well-oxygenated and poorly-oxygenated cancers. In a complementary approach, I will look to combine these predictions and other molecular information to identify those patients who would most benefit from specific therapies. If successful, this work might allow us to personalize therapy. For example, we could target toxic treatments of prostate cancer to only those individuals who might derive benefit, or we could identify those cervix cancer patients who have the highest risk of recurrence and whose outcome might be improved by more intensive therapy.