New AI-Based Imaging Method Can Detect Changes in Heart Function

Researchers from Rutgers Robert Wood Johnson Medical School (Rutgers RWJMS) and Robert Wood Johnson University Hospital (RWJUH) found that a new ultrasonic imaging method based on artificial intelligence (AI) techniques can detect very subtle changes in heart function. 

Located in New Jersey, Rutgers RWJMS includes 20 basic science and clinical departments and hosts numerous institutes, such as the Cardiovascular Institute and the Child Health Institute of New Jersey. RWJUH is a 614-bed facility with several centers of excellence, including cancer care, stroke care, neuroscience, and orthopedics.

In a new paper, researchers from both organizations described how they established new biological markers for cardiovascular disease using a new AI-based imaging method. These indicators can help clinicians detect cardiac issues earlier, allowing more time to make decisions regarding treatment efforts.

The paper published in the December edition of the Journals of the American College of Cardiology focused on how the researchers used AI modeling techniques to analyze echocardiogram images to develop precise interpretations of cardiac patterns that led to the detection of heart failure. A mouse model for heart failure helped researchers discover that the patterns are connected to microscopic changes in the heart muscle, a press release notes.

“By establishing and analyzing patterns of pixels obtained from the sample echocardiogram images, we were able to predict presence of heart conditions that can cause heart failure,” said Partho Sengupta, MD, the Henry Rutgers professor of cardiology and chief of the Division of Cardiovascular Disease and Hypertension at RWJMS, and chief of cardiology at RWJUH, in the press release. “Identifying changes in the heart muscle or cardiovascular function earlier can lead to more proactive interventions and the prevention of serious complications.”

The analysis conducted by Sengupta and his research partner, Naveena Yanamala, PhD, director of artificial intelligence and data science in the Division of Cardiovascular Disease and Hypertension at Rutgers RWJMS and director of RWJUH’s Innovation Center, established new biological markers for cardiovascular disease. Sengupta also mentioned that the biomarker could be applied to any current cardiac ultrasound device.

“This has the potential to give more people access to in-depth, expert analysis in a broad range of settings, leading to faster intervention and prevention of serious cardiac disease,” said Sengupta.

The project aligns with the mission of the Center for Innovation that RWJUH established earlier this year through a partnership with RWJMS, according to the press release. The mission is centered on developing new technologies to address healthcare issues.

Increasingly, AI is being applied in clinical cardiac care.

A study from May described an AI algorithm that assessed data from an Apple Watch device to detect a weak heart pump, specifically a low ventricular ejection fraction.

Researchers adapted the AI algorithm from a 12-lead ECG algorithm and then tested it with participants from 46 states and 11 countries. They worked with the Mayo Clinic Center for Digital Health to create a smartphone app that 2,454 patients used, allowing researchers to collect 125,610 ECGs. They used the data from the ECG recordings to test the adapted algorithm.

Another study from March involved Mayo Clinic researchers creating an AI computer-based algorithm that used voice biomarkers to locate clogged arteries, which could help clinicians detect heart issues.

By identifying more than 80 voice recording features, the app took six that correlated with coronary artery disease and combined them to arrive at a single score. When the score was high, it led to a high hazard ratio, which was linked to a higher risk of coronary artery disease-related medical events.