Big data for small patients - Building "child-size" individual predictive models for life after childhood cancer

Many children with cancer have radiation treatment as part of their care. As for all cancer treatments, there is a risk of lasting side-effects such as learning problems and reduced growth. Research is needed to reduce such side-effects, which is particularly important for children because of their long life expectancy. Radiation treatment is planned to give maximal dose to the tumour and minimal doses to nearby healthy organs. However, even with the most advanced ways of giving radiation (e.g. using the new Proton Beam Therapy machine in Manchester) it will never be possible to avoid all healthy organs.

This fellowship will find which parts of healthy organs are particularly damaged by radiation ('the important regions'). This knowledge would be incredibly useful when planning radiation treatments, because it is often possible to spare the important regions of an organ close to the tumour but not the whole organ. Hence, finding these important regions would be a step toward allowing reduced side-effects in many children with cancer.

The cancer centre with the most and the best documented children's health data in the world is St Jude Children's Research Hospital in Memphis, Tennessee. Children treated with radiation at St Jude have a very detailed and complete follow-up, and their side effects are measured using the most up-to-date methods.

In this project, we will:

(1) Set up a joint data analysis structure to show that our new method can be used on St Jude's data; with this, we will discover, for example, regions of the brain where radiation causes the most learning problems.

(2) Measure the changes in organ size and shape between children of different ages and sizes. For this we will use images from St Jude patients as well as from 500 healthy children in the United States, aged from 6 months to 16 years that were scanned every 2 years (we have permission to use these data for research). This information will help us make our method even more precise and able to find smaller "important regions". We will also use those images to build models of growth of the organs of interest (e.g. language center, hormone glands) in children, which will be useful for researchers studying other childhood diseases.

(3) Develop new and better ways to measure side-effects, using all the follow-up information obtained about a child's health as they grow into adulthood. This will mean, for example, that we can use images showing the health of each child even though taken many years after treatment.

This will be the first project of this kind focused on understanding side effects in children with cancer. In the future, the results from this project will help doctors give 'smarter' radiation treatments, with fewer side-effects. The models of growing organs will also be useful for research in other childhood diseases.

Economic and social impact:
Every year in the UK, 1400 children under the age of 15 are diagnosed with cancer. Thanks to advances in treatment, 80% of them survive for at least 5 years. It is estimated that around 20,000 people in England and up to 500,000 European citizens alive today are survivors of childhood cancer, and these numbers will increase with time and advancing therapies. However, many survivors face serious and life-threatening long-term effects as a result of their treatment-related toxicities. It is estimated that most (80%) childhood cancer survivors have one or more chronic health conditions, and that many (~50%) will experience severe or life-threatening complications up to several decades after treatment. US research has demonstrated that, compared to the general population, adult survivors of childhood cancer are twice as likely to report being unable to work because of health issues.
The economic burden of the life-long side effects associated with having cancer and its necessary treatment is considerable. Moreover, at the societal level, it has been shown that the lifetime costs in terms of decreased productivity or lost productivity due to premature death are much higher than direct health costs. By furthering our understanding of which dose to healthy organs/substructures is responsible for long-term side effects, this proposal will contribute to optimising cancer treatment and reducing side-effects in children. This project will therefore translate into increased quality of life, and preserved productivity for children treated in the future and their families.

Clinical impact:
The outputs of this project will enable clinicians to generate "smarter" radiotherapy plans to make sure that radiation doses avoid the sensitive areas identified through this fellowship. Such plans will minimise the incidence of side effects in children with cancer without affecting tumour control. This means that children will have a much better life after cancer treatment. In cases where it is not possible to avoid giving dose to sensitive areas, the doctors will be better able to inform each patient about their chances of getting particular side effects. This will enable patients to make more informed treatment decisions, know what to expect following treatment, and allow directed tests and interventions during follow-up to reduce the impact of the side effects.