A new artificial intelligence system developed by researchers in the United States provides a more accurate prediction of the likelihood that a patient with a heart problem, especially a diagnosed or suspected coronary heart disease, will die in the coming years.
The researchers, led by Dr. Theo Pezel of Johns Hopkins University Hospital in Baltimore, made the announcement at the European Society of Cardiology ‘EuroEcho 2011 scientific conference. The machine learning system can predict with 76% accuracy which cardiovascular patient will live or have died after ten years, ie it correctly predicts for three out of four patients.
Unlike traditional methods based on clinical data (age, sex, smoking, hypertension, cholesterol, etc.), the new system that “scores” each patient, includes information from cardiac imaging examinations based on the CMR (Cardiovascular) technique. Magnetic Resonance).
The system was “trained” by analyzing data on almost 32,000 patients with a mean age of 64 years between 2008-2018, who were followed for an average of six years for deaths of any etiology (2,679 people or 8.4% died during the study). An algorithm was created that “reads” the profile of each patient and makes an assessment of his future risk of death.
Follow Skai.gr on Google News
and be the first to know all the news
.
