Scientists have developed a new artificial intelligence tool that can predict a person's risk of more than 1,000 diseases and forecast changes in their health up to a decade in advance. The generative AI model, named Delphi-2M, was built by experts from the European Molecular Biology Laboratory (EMBL), the German Cancer Research Centre and the University of Copenhagen.
The tool uses algorithmic concepts similar to those behind large language models (LLMs) and was trained on anonymised patient data from 400,000 people in the UK Biobank study and 1.9 million patients in the Danish national patient registry. It assesses the probability of whether and when someone may develop conditions such as cancer, diabetes, heart disease and respiratory disorders.
Delphi-2M works by looking at 'medical events' in a patient's history, such as past diagnoses, alongside lifestyle factors including obesity, smoking and alcohol consumption, as well as age and sex. Health risks are expressed as rates over time, similar to a weather forecast giving a 70% chance of rain.
The research, published in the journal Nature, demonstrates that generative AI can model human disease progression at scale. 'Medical events often follow predictable patterns,' said Tomas Fitzgerald, a staff scientist at EMBL's European Bioinformatics Institute. 'Our AI model learns those patterns and can forecast future health outcomes.'
Ewan Birney, EMBL's interim executive director, said patients could benefit from the tool within the next few years. 'You walk into the doctor's surgery and the clinician is very used to using these tools, and they are able to say: Here's four major risks that are in your future and here's two things you could do to really change that.' He added that unlike existing single-disease models such as Qrisk, Delphi-2M can 'do all diseases at once and over a long time period'.
The team reported that Delphi-2M's accuracy is comparable to existing single-disease models and that its generative nature allows sampling of synthetic future health trajectories, providing meaningful estimates of potential disease burden for up to 20 years. Prof Moritz Gerstung of the German Cancer Research Centre said: 'This is the beginning of a new way to understand human health and disease progression.'
