Modelling and treatment verification in proton therapy

Proton Therapy

Conventional radiotherapy uses high energy beams of X-rays, or photons, to destroy cancerous cells. Whilst modern radiotherapy techniques allow the radiation dose to accurately target the tumour, some radiation is deposited in surrounding healthy tissue. This leads to treatment-related side effects impacting on patients’ quality of life. The use of high energy protons has the potential to reduce this radiation dose to surrounding tissues which reduces the risk of side-effects. However, the use of proton therapy in the UK is currently restricted to very specific cancers, like ocular cancer.

One aspect of this project is to investigate mathematical modelling of proton therapy treatment planning, improving the likelihood of eradicating the tumour and reducing radiation dose to surrounding healthy tissue. With conventional photon radiotherapy, the location of dose delivery can be inferred through measuring the x-rays that exit through the patient. Proton therapy treatment verification is much more challenging, and currently not available, as the majority of the radiation dose, and the proton beam itself remains within the patient.

Inverse problem

One solution to this issue is to measure the small amount of radiation that does escape from the patient during proton treatment. This radiation is a result of nuclear interactions and de-excitations and is released as high energy prompt-gamma radiation.

Several research groups worldwide are developing detector systems that are able to measure this radiation. However, due to sparsity of data, measurement of the prompt-gamma radiation alone does not provide assurance that the intended radiation dose was delivered as intended. This project will address this issue by providing accurate methods that allow the radiation dose distribution to be derived from a measurement of the emitted prompt-gamma photons.

Multiple mathematical techniques

IMI Fellows Alex Cox, Tristan Pryer and IMI Director Andreas Kyprianou collaborate on this project to provide accurate methods based on numerical and stochastic analysis of the Boltzmann transport equation. By combining and innovating a mixture of model adaptive methods together with ultra-modern Monte-Carlo methods, the radiation dose distribution can be derived from a measurement of the emitted prompt-gamma photons via an inverse problem. The model adaptive methods are based on previous EPSRC-funded work of Pryer. The Monte-Carlo methods make a remarkable transfer of methodology from previous work of Cox and Kyprianou for the Boltzmann transport equation for nuclear fission in the context of reactor safety assessment, which was funded by EPSRC with additional financial support from the IMI.