By Zoe Papadakis From Newsmax

Artificial Intelligence is now better than doctors at discerning who will die or have a heart attack. The amazing advance is due to machine-learning algorithms similar to those used by Netflix and Spotify.

Researchers at the International Conference on Nuclear Cardiology and Cardiac CT on Sunday explained how an algorithm analyzing 85 variables in 950 patients was able to learn how imaging data interacts. Using this data, it then identified patterns correlating the variables to death and heart attack. Its predictions were more than 90% accurate.

Artificial intelligence is built upon machine learning, which we encounter everyday through smartphones, Google’s search engines and when we binge watch shows on Netflix. Behind the scenes are machine learning algorithms that are constantly adapting according to the individual user.

Tapping into that potential could open up countless doors in the medical field, says study author Dr. Luis Eduardo Juarez-Orozco, of the Turku PET Centre, Finland.

“These advances are far beyond what has been done in medicine, where we need to be cautious about how we evaluate risk and outcomes,” Juarez-Orozco explained. “We have the data, but we are not using it to its full potential yet.”

Currently, doctors make treatment decisions according to risk scores but these are based on just a handful of variables and are often only moderately accurate. This is where AI has an advantage over humans. Through repetition and adjustment, machine learning can digest large amounts of data and identify complex patterns that might not be evident to humans.

Researchers set out to prove this in a study of 950 patients with chest pain who underwent standard protocol to look for coronary artery disease. They had coronary computed tomography angiography (CCTA) scans which produced 58 pieces of data based upon the presence of coronary plaque, vessel narrowing, and calcification.

Those that suggested disease then underwent a positron emission tomography (PET) scan which produced 17 variables on blood flow. Ten clinical variables were obtained from medical records including sex, age, smoking and diabetes.

Within a six-year follow-up of the participants, there were 24 reported heart attacks and 49 deaths that arose from various causes. Using this data and the 85 variables obtained from the study, a machine learning algorithm called LogitBoost was able to find the best structure to predict who had a heart attack or died.

“The algorithm progressively learns from the data and after numerous rounds of analyses, it figures out the high dimensional patterns that should be used to efficiently identify patients who have the event,” said Juarez-Orozco, noting doctors were already gathering extensive information about their patients, which could be used to predict individual risk among patients through machine learning.

These findings could have profound implications on future treatment. The Centers for Disease Control and Prevention found that about 735,000 Americans have a heart attack each year. While eliminating risk factors is the first step in fighting heart disease, advances that could help predict it in advance could save lives.

“This should allow us to personalize treatment and ultimately lead to better outcomes for patients,” Juarez-Orozco said.

For more on this story go to: https://www.newsmax.com/health/health-news/death-ai-artificial-intelligence-medicine/2019/05/13/id/915833/?ns_mail_uid=6952f1f9-507d-4a20-8cc0-0a1db158d76e&ns_mail_job=DM27722_05142019&s=acs&dkt_nbr=010502ssswm3