Researchers led by ETH Zurich and the University Hospital Zurich are now reporting that artificial intelligence (AI) is better at classification than cardiologists. The Zurich University Hospital announced on Tuesday that clear criteria for distinguishing between these two cardiovascular diseases are still lacking.
In the medical journal “JAMA Cardiology”, the researchers are now reporting on a machine learning algorithm that automatically classifies cardiovascular diseases based on imaging data. Data from cardiac ultrasound examinations of around 450 patients who were suffering from either an acute myocardial infarction or “broken heart” syndrome were incorporated into the algorithm.
This heart failure, also known as “Takotsubo” syndrome, occurs primarily in women and usually follows emotional or physical stress. A temporary weakness in the heart muscle causes part of the heart to inflate unnaturally, making it resemble a Japanese squid-catching vessel, a “takotsubo”. The consequences are chest pains, shortness of breath and, in the worst case, heart attack and death.
The researchers point out that despite the good performance of the AI, further studies are required before it can be used clinically.
Source: Nachrichten
