Listen to the audio version of the article
Algorithms “allies” of doctors and patients also in the world of cardiology: from prevention to diagnosis and treatment. In the space of a few years we have gone from looking for the symptoms of a disease on a search engine to using machine learning algorithms to promptly identify a heart attack, recognize a “difficult” coronary stenosis, choose the most suitable treatment or procedure for a heart disease. The specialists of the Italian Society of Interventional Cardiology (Gise) are taking stock on the occasion of the 44th National Congress, in Milan from 3 to 6 October.
«We are in the midst of a revolution in interventional cardiology and artificial intelligence is leading the way – explains Giovanni Esposito, president of Gise and director of the UOC of Cardiology, Hemodynamics and UTIC of the Federico II University Hospital of Naples – From myocardial infarction from the diagnosis and treatment of coronary atherosclerotic disease to the planning and execution of structural interventional procedures and the development of interactive educational applications and tools to provide people with information on cardiovascular diseases, risk factors and preventive measures: there are many possible applications and there will be many more in the future.”
In Italy approximately 120 thousand people suffer a myocardial infarction every year. Of these, around 25 thousand die because they were not rescued in time. The timeliness of the diagnosis is therefore crucial and the ECG is a non-invasive test used to evaluate the electrical activity of the heart. But how can artificial intelligence contribute to the early diagnosis of acute myocardial infarction? «AI is able to identify the electrocardiographic alterations that occur in the case of acute coronary syndrome – adds Esposito – In particular, recent studies have shown that the use of deep learning models achieve good accuracy in the diagnosis of heart attack. These observations pave the way for the use of AI systems to support the activities of time-dependent networks.”
Machine learning then allows the reconstruction, interpretation and analysis of angiographic images or images obtained with intravascular imaging methods. This means having tools capable of providing increasingly detailed information on the characteristics of coronary lesions. Esposito continues: «Specific algorithms can detect functionally significant coronary stenosis. Applications are available that combine angiographic and echocardiographic images in the machine learning model, allowing interventional cardiologists to identify soft tissue-based structures. This can allow for more intelligent anatomical orientation, particularly for difficult procedures, and reduce fluoroscopy time, contrast usage, and total procedure time. The development of non-invasive methods for the identification of significant coronary stenoses is also revolutionary.”
AI algorithms can help improve the quality of images obtained with transesophageal echocardiography, computed tomography (CT) or magnetic resonance imaging (MRI), facilitate their visualization and interpretation, but can also be used for training and training interventional cardiologists with the simulation of complex structural procedures in a safe virtual environment.