- Cancer Research UK and partners commit £10 million to create new tools using AI and state-of-the-art analytics to advance cancer early detection and prevention.
- Over the next five years, the Cancer Data-Driven Detection programme will enable access to data, link datasets and develop new tools to predict cancer risk.
- The programme will harness vast quantities of data to increase the number of people diagnosed with cancer at its earliest stages.
- At the University of Sheffield, we are leading the development of adaptive multimodal AI models for cancer risk prediction as part of this project
Today (23rd January 2025), Cancer Research UK, the National Institute for Health and Care Research (NIHR) and the Engineering and Physical Sciences Research Council (EPSRC) are announcing £10 million to create the Cancer Data-Driven Detection programme.
The programme aims to access and link data from different sources - including health records, genomics, family history, demographics, and behavioural data - to develop advanced statistical models that help scientists accurately predict who is most likely to get cancer. Alongside this, the programme will develop powerful new tools which use AI to analyse the data and calculate an individual’s risk of cancer throughout their lifetime.
Building the Infrastructure
Over the next five years, the funding will build the infrastructure required to access and link these datasets, train new data scientists, create the algorithms behind the risk models and evaluate the algorithms and AI tools to ensure that they are giving accurate and clinically useful information about cancer risk. The scientific programme will be guided by partnerships with cancer patients, the public, clinical experts and industry, while addressing ethical and legal considerations to ensure that the models and tools work well in practice.
At the University of Sheffield, we are leading the development of adaptive multimodal AI models for cancer risk prediction. This includes equipping models with robust domain adaptation capabilities to ensure they remain accurate and reliable across diverse populations, healthcare settings, and time periods. We are also integrating knowledge graphs to combine multimodal data and domain knowledge, which will enable interpretable predictions and provide deeper insights into decision-making processes. These efforts will be supported by our PyKale open-source software library, ensuring that the tools we develop are FAIR-compliant, transparent, and accessible to the global research community.
Professor Haiping Lu, Professor of Machine Learning in the School of Computer Science and Head of AI Research Engineering at the Centre for Machine Intelligence at Sheffield, said: ““By developing adaptive multimodal AI models with robust domain adaptation and knowledge graph integration, we aim to ensure that cancer risk predictions are not only accurate but also interpretable and reliable across diverse populations and healthcare settings. Our commitment to open-source, FAIR-compliant software ensures that these innovations are accessible and impactful, enabling earlier and more precise detection of cancer to benefit both patients and healthcare systems globally.”
The models generated from this research could be used to help people at higher risk of cancer in different ways. For example, the NHS could offer more frequent cancer screening sessions or screening at a younger age to those at higher risk, whilst those at lower risk could be spared unnecessary tests. People identified as higher risk could also be sent for cancer testing faster when they go to their GP with possible cancer signs or symptoms. Individuals at higher risk could also access different ways to prevent cancer.
Finding those most at risk of cancer
Earlier diagnosis of cancer saves lives. Yet according to analysis of NHS figures by Cancer Research UK, only 54.4% of cancers in England are diagnosed at stages one and two*, where treatment is more likely to be successful. NHS England has set a target to diagnose 75% of cancers at stages one and two by 2028, and this will only be achieved with research and embracing new technologies to catch cancer earlier.
Head of Prevention and Early Detection Research at Cancer Research UK, Dr David Crosby, said: “The single most important thing we can do to beat cancer is to find it earlier, when treatment is more likely to be successful. With half a million cancer cases per year expected in the UK by 2040, we need a major shift towards more accurate diagnosis and detection of early cancer.
“The Cancer Data Driven Detection programme will link health data sources together and build the powerful new tools doctors need to identify those at highest risk of cancer and prioritise resources towards them. Moving towards a preventative approach to healthcare will not be easy and will take time. Cancer Research UK’s investment in the programme is an investment in the future of cancer care.”
Professor Lucy Chappell, Chief Scientific Adviser at the Department of Health and Social Care (DHSC) and Chief Executive Officer of the National Institute for Health and Care Research (NIHR) said: “Detecting and diagnosing cancer earlier is key to improved survival and quality of life for patients. By leveraging AI to enable healthcare professionals to identify people at a greater risk of cancer, this initiative could improve the way patients are screened and diagnosed. This programme’s AI-driven insights could lead to more effective treatment and improved survival, helping patients to live longer, healthier lives”
Executive Director of Research at EPSRC, Jane Nicholson, said: “Earlier detection of cancer can provide a wider range of treatment options and save lives. This programme, leveraging the power of artificial intelligence and datasets, could revolutionise early detection and intervention.
“By working with a wide range of stakeholders, including cancer patients, the models and tools developed through the Programme could equip individuals with the crucial knowledge about their potential for developing cancer.”
Science Minister, Lord Vallance, said: “There are huge opportunities in AI to improve UK healthcare, from scans detecting illnesses earlier to bringing NHS waiting lists down by planning appointments more efficiently, and these will continue to develop.
“This investment in harnessing the potential of data to spot those at risk of cancer represents the sort of innovation the Government’s new AI Opportunities Action Plan sets out to realise, so this technology improves lives, while transforming public services and boosting growth.”
The Cancer Data Driven Detection programme is jointly supported by Cancer Research UK, the National Institute for Health & Care Research, the Engineering and Physical Sciences Research Council, Health Data Research UK, and Administrative Data Research UK.