The use of artificial intelligence (AI) to evaluate and treat heart disease has been one of the biggest stories at the American Heart Association’s Scientific Sessions 2023 conference in Philadelphia.

After years of hype about AI, and some significant breakthroughs along the way, is it finally time for more and more cardiologists to start using these advanced algorithms on a daily basis?

“Computational methods to develop novel predictors of health and disease—‘artificial intelligence’—are becoming increasingly sophisticated,” Dan Roden, MD, a professor with Vanderbilt University Medical Center and chair of the AHA’s Council on Genomic and Precision Medicine, said ahead of the three-day event.

These are just some of the big AI studies presented during the first few days of AHA 2023:

1. AI shows potential to accelerate heart attack diagnosis, treatment

According to new data out of Taiwan, advanced AI can help shave nearly 10 minutes off of the total time needed to diagnose and treat a ST-elevation myocardial infarction (STEMI). The ARISE trial focused on more than 43,000 patients. AI-powered electrocardiogram (ECG) evaluations were used to evaluate approximately one-half of the study’s patients, and the others were treated without any AI assistance. The full abstract was presented during AHA 2023.

The AI-ECG algorithm used in the analysis achieved a positive predictive value of 88%, a negative predictive value of 99.9% and helped reduce the median time from ECG to cath lab from 52.3 minutes to 43.3 minutes.

“Hospitals can use AI tools more to help front-line doctors, especially those with less experience,” lead author Chin-Sheng Lin, MD, PhD, vice dean at the School of Medicine at the National Defense Medical Center in Taiwan, said in a statement. “This could lead to faster treatment and less mistakes when it comes to treating patients who are experiencing heart attacks.”

Lin also emphasized that the algorithm was proven to be especially helpful when treating hospitalized patients, calling its impact “astonishing.”

“This tells us that there's a lot we can do to improve how we diagnose STEMI in hospitalized patients,” he said.

2. Smartphone-based AI monitors voices of heart failure patients for signs of trouble

Another abstract presented at AHA 2023 explored AI’s ability to detect key differences in a heart failure patient’s voice that could suggest they face an increased risk of hospitalization. The study focused on Cordio Medical’s smartphone-based speech analysis software, the HearO system. HearO was designed to evaluates the pitch, volume, dynamics and other characteristics of a patient’s speech pattern, alerting specialists when it believes worse heart failure symptoms could be on the horizon. The algorithm built into HearO was trained using data from 263 patients and then validated with data from another 153 patients.

Overall, the application was able to predict worsening heart failure symptoms in 76% of patients nearly a month before hospitalization. The AI offering did generate an average of three unnecessary alerts per year for each patient, and the study included a relatively low number of patients, but researchers were encouraged by these findings.

“Speech analysis is novel technology that may be a useful tool in remote monitoring of heart failure patients, providing early warning of worsening heart failure that frequently results in hospitalization,” lead author William T. Abraham, MD, a professor with The Ohio State University Wexner Medical Center, said in a statement. “This technology has the potential to improve patient outcomes, keeping patients well and out of the hospital, through the implementation of proactive, outpatient care in response to voice changes.”

3. AI can reduce number of false alerts from insertable cardiac monitors

Advanced AI models can make a significant impact on the number of false alerts associated with insertable cardiac monitors (ICMs), according to one abstract presented at AHA 2023. The research focused on Medtronic’s LINQ II ICM, an implantable device approved by the U.S. Food and Drug Administration for multiple indications. Medtronic wanted the study to measure the impact of using its AccuRhythm AI algorithm enhancements to improve the performance of the LINQ II ICM.

Overall, researchers determined that the AccuRhythm algorithms helped reduce false atrial fibrillation alerts associated with the LINQ II by more than 91%. According to Medtronic, this could save physicians more than 400 hours of time over the course of a single year.

“What this validation dataset suggests is that AI algorithms can help device clinics better manage their ICM patients, saving time and minimizing data review burden,” lead author Jagmeet Singh, MD, a professor of medicine at Harvard Medical School and Resynchronization and Advanced Cardiac Therapeutics Program at the Massachusetts General Hospital Heart Center, said in a statement from Medtronic. “When we apply AI, we observed that on average, a LINQ II ICM patient will now have about one false, or non-clinically relevant alert per year related to AF or pause, which is a manageable number for most clinics.”