Every year, 10 million people die from cancer, and this number continues to rise particularly among young adults. Alarmingly, half of these deaths are avoidable through prevention, earlier detection, or timely treatment. And other chronic diseases such as neurological, metabolic, and cardiovascular also suffer from the same missed opportunities. The main problem is that today’s systems are reactive, focusing on treating diseases once they appear (“sick-care”) rather than proactively maintaining health (“health-care”).
To combat this, we need to develop tools to accurately predict disease risk years ahead. For example, colon cancer screening typically begins at age 45. However, younger individuals are also at risk, like the actor Chadwick Boseman, who was diagnosed with Stage 3 colon cancer at 39 and passed away at 43. If he had been aware of his disease risks earlier in life, he could have taken preventive measures, such as undergoing earlier screenings and making lifestyle changes to reduce those risks.
Current solutions fall short by focusing on a limited set of biomarkers (traditional blood tests) or relying solely on genetic predispositions without considering lifestyle factors (DNA tests). However, emerging technologies now allow us to measure tens of thousands of molecules in the blood, offering a comprehensive snapshot of both our current and future health.
The UK Biobank is an invaluable resource that has collected extensive data from hundreds of thousands of individuals over time. This includes thousands of blood measurements using technologies from companies like Olink and Nightingale, along with detailed information on lifestyle and environmental factors that influence health outcomes.
Our project aims to harness this data-including DNA, proteins, metabolites, and lifestyle factors-to develop a comprehensive disease risk assessment model for conditions such as cancer, neurological, cardiovascular, and metabolic diseases. By analyzing more than 3 billion data points per individual and integrating this information with the existing biomedical literature, we aim to create detailed profiles of disease risks and provide actionable prevention plans.
This project will take 36 months. Initially, we will meticulously clean the data to avoid the “garbage in, garbage out” effect. Next, we will identify which diseases have sufficient data for reliable risk estimation, allowing us to train and validate our AI models. Finally, we will share our findings through conference presentations and scientific publications, and seek clinical and industrial partners to translate our research into practical and accessible tools such as a blood test and a lifestyle questionnaire.