AI better than doctors at predicting when hearts will failBy Staff Writer 17 January 2017 | Categories: news
As we’re starting to find out, computers are better than us at a number of things, and this now includes detecting when your heart will fail. According to the BBC, artificial intelligence software can correctly predict 80% of the time which patients with pulmonary hypertension would still be alive after one year. For doctors, this figure stands at 60%.
The team behind this research, from the UK's Medical Research Council, believes that the software would be able to help by correctly spotting those patients who require more aggressive treatment.
Pulmonary hypertension (high blood pressure in the lungs) does damage to parts of the heart, with a number of treatments available, including lung transplants and medication. Doctors, though, need to know how long a patient will live in order to provide the correct treatment, and this is where the software comes in.
Using machine learning capabilities the software was able to study MRI scans of 256 patients’ hearts, plus their blood test results, with it measuring 30 000 different structural points during heartbeats. When put together with eight years of patient health records, the artificial intelligence software could learn which abnormalities predicted when patients would die.
While this spells better treatment abilities, the team noted that it would like to test it on more patients first, before making it available to doctors. The research was published in the journal Radiology.
It is believed that artificial intelligence will play a major role in medicine in years to come. Recently, for example, IBM’s Watson, a supercomputer with artificial intelligence capabilities, correctly identified a particular variation of leukaemia to help treat a Japanese patient – a diagnosis doctors could not get right.
Most Read Articles
Have Your Say
What new tech or developments are you most anticipating this year?