Abstract
**The American Heart Association (AHA)** annual scientific sessions have always been the largest international cardiovascular scientific event closing out each calendar year. AHA 2023, held in Philadelphia this past November, devoted considerable time to presenting, sharing, and discussing artificial intelligence (AI) applications at the bedside, featuring several clinical trials presented for the first time, including: (1) the SPEC-AI1 trial, screening nearly 1,200 pregnant and postpartum women in Nigeria using a digital stethoscope, which doubled the number of newly detected heart failure cases (which would otherwise have been missed under usual care), enabling timely appropriate management and reducing mortality or complications; (2) the SPEECH trial in 400 adults with heart failure in Israel, leveraging smartphone-based voice analysis technology to detect early signs of heart failure deterioration, achieving 71% accuracy in detecting heart failure events approximately 3 weeks before hospitalization; (3) the ARISE trial in patients with ST-elevation myocardial infarction (STEMI) at a hospital in Taiwan, applying AI-assisted ECG analysis to enhance diagnostic capability (positive predictive value of 88% and negative predictive value of 99.9%), enabling STEMI patients to be transferred to the catheterization laboratory nearly 10 minutes earlier; and (4) the ORFAN2 trial, evaluating additional inflammatory risk based on pericoronary adipose tissue on routine multi-slice coronary CT angiography, which increased the ability to predict cardiovascular events independently of clinical risk scores (reclassifying approximately 30% of patients into a higher-risk group and about 10% into a lower-risk group). The number of clinical trials involving medical devices using artificial intelligence or machine learning (AI/ML) has been increasing in recent years: from 1 study in 2010 to 619 studies in 2022. The World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) recorded a total of 2,669 trials registered during the 2010–2023 period, mainly in specialties such as diagnostic imaging (27%), general hospital management (13%), gastroenterology and urology (12%), neurology (10%), and cardiology (9%). However, the majority of these studies remained limited to a national scale (97%), led by China (1,095 studies), the United States (196), Japan (162), and India (139). In the second half of 2023, for the first time, the U.S. Food and Drug Administration (FDA) approved 2 artificial intelligence algorithms for use as cardiovascular screening tools, both utilizing ECG data: one to identify heart failure with reduced ejection fraction (the Anumana ECG-AI model, FDA-approved October 2023), and the other to identify hypertrophic cardiomyopathy (the Viz.AI model, FDA-approved August 2023). This approval marks a milestone in which artificial intelligence application models are truly recognized as an independent category of medical device—software as a medical device (SaMD)—competing directly with other diagnostic and therapeutic modalities (drugs, biologics, medical devices, etc.) in the field of cardiology. Compounding the buzz surrounding ChatGPT and OpenAI in the final months of 2023, people are beginning to sense that artificial intelligence has arrived very close at hand…
