A new step has been taken in medicine — scientists have developed an artificial intelligence model capable of predicting diseases that may appear in the human body in the future. The program, called Delphi-2M, analyzes medical data and can calculate the risk of more than a thousand diseases. According to scientists, it can identify possible health problems even ten years in advance.
How does the model work?
Delphi-2M was trained on anonymous medical records. It analyzes hospital records, doctors’ notes, and patients’ daily habits. For example, smoking or lifestyle factors also play an important role in its calculations.
The program works like a weather forecast: just as it is said “it may rain tomorrow,” Delphi-2M concludes “there is a probability of such a disease occurring.”
Testing and trial results
Initially, the model was trained on data from more than 400,000 participants in the UK Biobank project. Later, it was tested on the medical records of 1.9 million people in Denmark. The results showed that the program’s predictions matched reality.
Delphi-2M proved especially effective at predicting diabetes, heart attacks, and sepsis. However, it has limitations in forecasting random infections.
Importance for medicine
Experts emphasize that such artificial intelligence will provide great assistance to doctors in the future. If a patient shows a predisposition to a particular disease, the doctor may recommend lifestyle changes or prescribe medication in advance. For example, someone at risk of liver disease would be advised to avoid alcohol.
Moreover, this technology has strategic importance for the entire healthcare system. It makes it possible to calculate in advance how many patients may appear in a particular region and to plan hospital capacity accordingly.
Scientists’ opinion
Professor Moritz Gerstung, head of the artificial intelligence department at the German Cancer Research Center (DKFZ), called this program “the beginning of a new era in understanding human health.” According to him, in the future, models like Delphi-2M will help create individualized treatment plans for each patient.
How will it be further developed?
So far, the model has been trained mainly on data from people aged 40–70, so its capacity to analyze health in other age groups is limited. Scientists plan to improve it by including additional data — blood tests, genetic information, and imaging.
The full results of the research were published in Nature. The study was carried out in collaboration with the European Molecular Biology Laboratory, the German Cancer Research Center (DKFZ), and the University of Copenhagen.
Read 'Zamin' on Telegram!
