Development and Assessment of a Genetic and Environmental Risk Model for Colorectal Cancer

Bowel cancer is the second most common cause of death from cancer in the UK. It is often preventable through screening programmes, as it usually starts as a non-cancerous growth (a polyp) which can be removed during an examination of the large bowel using a colonoscope. Colonoscopy is the best screening method, but is time-consuming, expensive and may occasionally produce complications, and so other tests looking for blood in the stool can be used to identify people more likely to have polyps or cancer. They will then have a colonoscopy. However, in the UK 50% of those with evidence of blood in the stool do not have cancer or polyps at colonoscopy, and hence better screening methods are needed.

Many factors increase an individual's risk of developing bowel cancer. These include personal or 'environmental' factors such as age, gender, medical history, and diet, and our genes. Certain common genetic variations (single nucleotide polymorphisms, SNPs) in the code making up our genes are known to moderately increase the risk of bowel cancer, and in combination, these variations can have a big effect. Some rare families have a very high inherited risk of bowel cancer because of a change in one of a small number of high risk genes. These account for around 5% of cases of bowel cancer.

The project aims to develop a way of calculating a person's risk of developing bowel cancer by looking at their genetic and environmental/personal risk factors. Incorporating risk information would inform screening decisions, potentially avoiding over-investigation of those at low risk, improving health outcomes, and allowing more effective use of NHS resources. Furthermore, screening for the rarer changes in high risk genes which cause bowel cancer can identify individuals and families with these genes, allowing them to have appropriate intensive screening, and other interventions to reduce their chances of developing cancer and other associated problems.

The project will be conducted in collaboration with researchers across several departments at the University of Oxford, in addition to colleagues at the Institute of Cancer Research (London) and the University of Amsterdam.

I will begin by analysing data from around 23,000 people with bowel cancer to identify further SNPs that increase bowel cancer risk; around 40 SNPs are already known. I will combine data from several previous studies with a large number of newly tested samples; in other cancers (such as breast cancer) many more SNPs have been found by looking at larger groups of patients. This information will contribute to our knowledge of causes of bowel cancer, and improve the accuracy of the colorectal cancer risk model I will then develop. This model will be created by comparing information on risk factors of individuals with and without colorectal cancer in the UK Biobank study. Several different versions of the model will be developed to find the most accurate. If possible the final model will then be tested in other populations.

Finally I will develop the technology for performing genetic testing which will work reliably at the scale needed for a national bowel cancer screening programme. This will involve identifying and optimising an appropriate saliva based test kit, and the technology to detect the gene changes in each sample, including computer programmes to analyse the data. I will also assess the costs of these processes to provide information on the resources required to incorporate the model into the bowel screening programme.

The output will be a risk model predicting a person's risk of colorectal cancer which could be used to personalise their bowel cancer screening, and the technology needed to do this. Beyond this project, this risk model and technology will be assessed for cost-effectiveness and trialled in individuals undertaking bowel screening, aiming to incorporate this into national bowel cancer screening programmes to improve patient outcomes.

Colorectal cancer (CRC) is the second leading cause of cancer deaths in developed nations. Whilst CRC is highly preventable using colonoscopic screening, this is expensive. Intermediate tests such as faecal occult blood tests (FOBT) are therefore used to identify individuals for further investigation. Modulating screening modality, interval, test interpretation and follow-up by prior risk could enable better targetting of NHS funding, and improve outcome measures such as cancer incidence and quality of life. Genetic risk is a major contributing factor, including around 40 common polymorphisms with moderate effects and 14 rare high risk genes. Non-genetic factors such as age, family and personal medical history, and diet also affect CRC risk.

I will initially use a genome wide association study meta-analysis of state-of-the-art data to identify more common CRC single nucleotide polymorphisms (SNPs), followed by additional genotyping to identify further new CRC SNPs. I will then profile the known genetic and non-genetic factors in a nested case-control study within UK Biobank to construct and quantitate a CRC risk model, initially using classical logistical regression based approaches. This will then be validated in other cohorts. An appropriate genetic testing methodolgy for bowel cancer screening programmes (based on saliva samples) will be identified and optimised using an amplicon based gene and SNP panel and next generation sequencing platform. I will develop a bioinformatics pipeline for analysis of this data. Finally, I will summarise the resources required to use this new technology and associated costs, to provide information on the budget impact for the NHS.The project will give new insight into cancer risk, and provide a genetic and environmental risk model and associated technology. These have the potential to be applied nationally and internationally to bowel cancer screening programmes, thereby improving patient outcomes and resource management.

This identification of new risk single nucleotide polymorphisms (SNPs) for colorectal cancer (CRC) and development and assessment of a new risk profiling model for use with bowel cancer screening programmes will benefit a wide range of people and organisations.

Academic impact:
Academics within the field of CRC genetics, genetic epidemiology, screening policy and population health will benefit from the research. Identification of new risk SNPs will further our understanding of population risk and might highlight new pathways for functional analysis, potentially in the long term leading to identification of therapeutic targets. The development of a comprehensive risk model combining genetic and environmental/lifestyle data will provide a platform from which models for other population groups can be developed. In addition, the amplicon-based sequencing platform developed for analysing both Mendelian genes and SNPs will be of use to other research groups.

UK population and NHS impact:
The project has the potential to positively impact on the NHS and the health of the UK population as a whole. Risk stratification would prevent over-investigation of low risk individuals, targeting resources to the appropriate investigation of those at higher risk. It could also help to resolve current management issues within bowel cancer screening, for example how to manage those with repeatedly positive stool testing but normal colonoscopies. The development of this colorectal cancer risk model, is the first step towards personalising bowel cancer screening in this way. Risk stratification could also be used to guide chemoprevention strategies. Additionally the detection of individuals carrying Mendelian genes will have a significant impact on their health, allowing enhanced screening, chemoprophylaxis, and prophylactic surgery to be performed and avoiding the burden of cancer treatment to both the patient and NHS. This would benefit additional family members, in whom genetic testing and appropriate management could be initiated.

Following this project, the cost-effectiveness of incorporating the risk model and impact on bowel screening programmes would be assessed through economic modelling. We would aim to trial the model in a screening population in collaboration with the NHS Bowel Cancer Screening Programme, and international screening programmes (initially through collaboration with colleagues in the Netherlands). The ultimate aim of the project is to improve patient care and quality of life across the range of individuals involved in the screening programme, in a cost-effective manner.

Genetics policy makers:
This work will demonstrate the utility of population level testing for both moderate risk variants and Mendelian conditions. As such it will contribute to growing interest and evidence for the incorporation of genetic information into routine healthcare, and inform national funding and policy decisions in the future.

International impact:
The CRC risk model would be applicable to populations of European origin, and thus has the potential to improve screening programmes and population health in Europe, North America, and Australasia. Through collaboration with Professor Dekker in Amsterdam, we would plan to demonstrate this in the Dutch population. Thus this research would hopefully have a positive impact on population health on an international scale.

Timescale:
Results of the GWAS meta-analysis would be available within the first 2 years of the DPhil programme, with further study results within the 3 years of the programme. Data will also be returned to UK Biobank within this time. Beyond this we would aim to trial the risk model within a bowel screening population within 1-2 years following the completion of the project. The model could potentially be incorporated into national bowel cancer screening programmes, both within the UK and abroad, within 5 years of completion of this project.