Cancer causes more than 1 in 4 of all deaths in the UK, with around 167,000 lives lost to the disease each year. While there have been dramatic improvements in cancer treatment over recent decades causing the total number of cancer deaths to decline by about 12% since the 1990s, the mortality for several types of cancer continues to increase, in some cases sharply. Screening and detection of potentially lethal cancers at earlier stages of the disease holds the key for effective cancer treatment. Selection for screening is currently done based on simple variables such as age, which is a major cancer risk factor, and current screening programmes are limited to a handful of specific cancer types. We propose to develop and validate new models for predicting the risk of developing multiple types of cancer. By integrating various types of data that go far beyond single variables like age-such as genetic information, blood markers, imaging results, and clinical data-these models will aim to provide more accurate assessments of an individual’s cancer risk across different types of cancer. The UK Biobank, with its vast repository of health-related data, provides a unique opportunity to develop these advanced risk prediction models. Using machine learning techniques, we will analyse this rich dataset to identify genetic factors, biomarkers, and other variables associated with cancer risk. We will then use this information to construct predictive models capable of estimating an individual’s likelihood of developing various types of cancer over time. The planned duration of the project is 36 months. The ultimate goal of this research is to improve cancer screening and early detection strategies by tailoring them to each person’s specific risk profile. By accurately identifying individuals at higher risk of developing cancer, healthcare providers can offer more focused risk-appropriate screening and monitoring protocols, potentially catching cancer at earlier, more treatable stages. This approach promises to improve patient outcomes by facilitating timely interventions in the future.
