Prevention is the best remedy against diseases, and when it is done before, better. Thanks to artificial intelligence, we can already Prevent more than 1,000 diseases up to 20 years before they occur.
He German Cancer Research Center (DKFZ) has carried out a proof of concept, consisting of train a Language model generative with anonymized medical data of 400,000 patients from the United Kingdom bioban. These data were crossed with another 1.9 million patients from the National Registry of Patients in Denmark.
The result is an artificial intelligence that can identify more than 1,000 future diseases simply by examining the medical history and medical reviews of a patient.
The AI that predicts diseases
Logically, there are certain types of ailments that stay out, such as infectious, since they depend on contact with viruses or bacteria, and that cannot be predicted.
However, it can «intuit» diseases such as heart attacksall types of cancer, diabetesand genetic diseasesyears in advance, and probabilities of up to 70%.
The press release clarifies that this is similar to predicting time. The AI offers an approach, but does not mean that it will be successful 100%. However, being a forecast of years, even two decades, even serious diseases can be eliminated, or minimized its impact, with prevention treatments.
«By modeling how diseases develop over time, we can begin to explore when certain risks arise and what is the best way to plan early interventions. It is a great step towards more personalized and preventive approaches to health care,» explains Dr. Ewan Birney, one of the authors of the project.
And how does this work that predicts diseases? «Like the large linguistic models can learn the structure of the phrases, this model of the ‘grammar’ of the health data to model medical records as sequences of events that take place over time,» explains Moritz Gerstung, of the DKFZ.
«These events include medical diagnoses or lifestyle related factors, such as smoking.» The model learns to forecast the risk of disease from the order in which these events occur and the time between them.
A revolution in medical prevention
«Medical events usually follow predictable patterns,» acknowledges Tom Fitzgerald, of the European Bioinformatics Institute (EMBL-EBI). «This model of the AI learns these patterns and can predict future health results. It allows us to explore what could happen based on the medical history of a person and other key factors.»
At the moment, this artificial intelligence that predicts diseases It is not going to be used with real patients, because training databases, although very wide, are not complete. For example, most patients included are between 40 and 60 years old, so it does not work for young people or the elderly.
But the concept of concept has turned out to be a success, opening the doors to disease treatment even before they occurcompletely personalized.
Know How we work in NoticiasVE.
Tags: Artificial intelligence, health, viral