Oxford unveils AI capable of predicting heart attacks 10 years in advance

by

in ,

With the explosion of AI, it continues to push the boundaries of medicine and pharmaceutical research ever further. Diagnostic support, therapeutic decision support, development of new drugs or personalized therapies, etc. This time, it is researchers from the prestigious University of Oxford who have produced a real miracle: an AI-based heart scanning system to detect heart attack risks for years before they manifest themselves.

Unparalleled precision

Every year, we see an influx of practitioners from the United Kingdom more than 300,000 people complaining of chest pain. Most of them are then forced to undergo examinations to detect possible cardiac abnormalities.

However, these examinations only too rarely reveal cardiac obstructions, since less than one patient in five actually presents this type of lesion. The vast majority of patients are therefore sent home, with the false assurance that the heart is healthy. A doctor interviewed by The Guardian told the newspaper that around 80% of patients return home with the false impression of being safe from any danger.

It is in this context that Oxford researchers have designed a most innovative technology, which aims to fill these diagnostic gaps by analyzing the damage caused by inflammation of the arteries, lesions often obscured by traditional examinations.

This AI system thus offers a much more precise and reliable vision of the risks incurred by patients suffering from cardiac pathologies. According to its designers, this new approach could save thousands of lives every year. Also, NICE (National Institute for Health and Care Excellence) and the NHS (National Health Service) should decide on this innovation by the end of the current year. To support these statements, recent studies demonstrate that after consulting the data provided by this artificial intelligence, doctors adjusted their therapeutic management in nearly 45% of cases.

This discovery, more than promising, suffers from the same pitfalls as all AI systems: their propensity to “hallucinate”. The risk of misinterpretation of data exists and would obviously not have the same consequences as a certain ChatGPT which would provide erroneous information. However, the first results still remain very encouraging and it will be necessary integrate robust enough controls to make the system completely reliable.

  • An AI system developed by Oxford researchers can predict heart attacks years in advance by analyzing inflammation in the arteries.
  • This could save thousands of lives, with preliminary tests having shown therapeutic adjustments in nearly 45% of cases after its use.
  • To ensure reliable diagnostics, however, security controls will need to be put in place.

📍 To not miss any news from Presse-citron, follow us on Google News And WhatsApp.