Presentation on YouTube.

In this talk I will discuss two applications of mathematical modelling in oncology.

The first concerns the detection of quantifying response to drug treatment in mouse models of human cancer. To be able to translate the results of the experiments more readily, a preferred measure to quantify this response should take into account more of the available experimental data, including both tumour size over time and the variation among replicates. I will present and discuss a new measure, KuLGaP, that is based on Gaussian Process regression. Our results show that KuLGaP is more selective than currently existing measures, reduces the risk of false-positive calls, and improves translation of the laboratory results to clinical practice.

The second application concerns intensity modulated radiotherapy (IMRT), a form of cancer treatment that uses ionising radiation. To minimise the impact of errors in patient positioning, treatment plans are designed using robust optimisation. Since some parts of the radiation will hit surrounding tissues, we are necessarily faced by a multicriteria problem: optimise the radiation dose delivered to the tumour while minimising the dose delivered to adjacent organs. Changing the weights of each objective will give different solutions, and this adjustment is often done manually, with input from doctors. In this talk I will present a novel algorithm, inspired by ideas from machine learning, that guides this process and provides doctors with new treatment plans. Rather than displaying a single solution, multiple possible plans are offered and represented visually for radiologists and oncologists to choose from.