Newswise – A special artificial intelligence (AI) -based computer algorithm created by researchers at Mount Sinai was able to learn to identify subtle changes in EKGs (also called ECGs or ECGs) to predict whether a patient had heart failure .
“We have shown that deep learning algorithms can recognize problems with pumping blood on both sides of the heart from ECG waveform data,” said Benjamin S. Glicksberg, PhD, assistant professor of genetics and genomic sciences, member of the Hasso Plattner Institute for the digital. Health at Mount Sinai, and lead author of the study published in the Journal of the American College of Cardiology: Cardiovascular Imaging. “Usually, the diagnosis of this type of heart disease requires expensive and time-consuming procedures. We hope that this algorithm will allow a faster diagnosis of heart failure.
The study was led by Akhil Vaid, MD, a postdoctoral researcher who both works in the Glicksberg Girish N. Nadkarni, MD, MPH, CPH, associate professor of medicine at the Icahn School of Medicine in Mount Sinai, head of the Data-Based and Digital Medicine (D3M) division, and lead author of the study.
Affecting approximately 6.2 million Americans, heart failure or congestive heart failure occurs when the heart pumps less blood than the body normally needs. For years, doctors have relied heavily on an imaging technique called an echocardiogram to assess whether a patient has heart failure. Although useful, echocardiograms can be a labor-intensive procedure that is only offered in certain hospitals.
However, recent breakthroughs in artificial intelligence suggest that EKGs, a widely used electrical recording device, could be a quick and readily available alternative in these cases. For example, many studies have shown how a “deep learning” algorithm can detect weakness in the left ventricle of the heart, which expels freshly oxygenated blood to the rest of the body. In this study, the researchers described the development of an algorithm that evaluates not only the strength of the left ventricle but also the right ventricle, which takes deoxygenated blood from the body and pumps it to the lungs.
“Although attractive, it has always been difficult for doctors to use ECGs to diagnose heart failure. This is in part because there are no established diagnostic criteria for these evaluations and because some changes in ECG readings are just too subtle for the human eye to detect, ”said the Dr Nadkarni. “This study represents an exciting step forward in finding information hidden in ECG data that may lead to better screening and treatment paradigms using a relatively simple and widely available test.”
Typically, an EKG involves a two-step process. Conductive wires are taped to different parts of a patient’s chest, and within minutes a specially designed portable machine prints a series of wavy lines, or waveforms, representing the electrical activity of the heart. These machines can be found in most hospitals and ambulances across the United States and require minimal training to operate.
For this study, the researchers programmed a computer to read the patients’ EKGs as well as data extracted from written reports summarizing the results of corresponding echocardiograms taken from the same patients. In this situation, the written reports acted as a standard set of data that the computer could compare with the EKG data and learn to spot weaker hearts.
Natural language processing programs have helped the computer extract data from written reports. Meanwhile, special neural networks capable of discovering patterns in images have been incorporated to help the algorithm learn to recognize pumping forces.
“We wanted to push the state of the art by developing AI that can understand the whole heart easily and inexpensively,” said Dr Vaid.
The computer then read over 700,000 EKGs and echocardiogram reports obtained from 150,000 patients in the Mount Sinai health system from 2003 to 2020. Data from four hospitals was used to train the computer, while the data from a fifth was used to test the performance of the algorithm. in a different experimental setting.
“A potential benefit of this study is that it involved one of the largest collections of ECGs from one of the most diverse patient populations in the world,” said Dr Nadkarni.
The first results suggest that the algorithm was effective in predicting which patients would have healthy or very weak left ventricles. Here, force was defined as the left ventricular ejection fraction, an estimate of the amount of fluid the ventricle pumps with each beat, as seen on echocardiograms. Healthy hearts have an ejection fraction of 50% or more, while weak hearts have an ejection fraction of 40% or less.
The algorithm was 94% accurate in predicting which patients had healthy ejection fraction and 87% accurate in predicting those with less than 40% ejection fraction.
However, the algorithm was not as good at predicting which patients would have slightly weakened hearts. In this case, the program was 73 percent accurate in predicting patients who had an ejection fraction between 40 and 50 percent.
Other results suggest that the algorithm also learned to detect right valve weaknesses from EKGs. In this case, weakness was defined by more descriptive terms taken from echocardiography reports. Here, the algorithm was 84 percent accurate in predicting which patients had weak right valves.
“Our results suggest that this algorithm could potentially help doctors correctly diagnose heart failure on either side of the heart,” said Dr Vaid.
Finally, further analysis suggested that the algorithm may be effective in detecting cardiac weakness in all patients, regardless of race and gender.
“Our results suggest that this algorithm could be a useful tool to help clinicians deal with heart failure in a variety of patients,” added Dr. Glicksberg. “We are carefully designing prospective trials to test its effectiveness in a more real-world setting. ”
This study was supported by the National Institutes of Health (TR001433).
Vaid, A. et al., Using Deep Learning Algorithms to Identify Right and Left Ventricular Dysfunction Simultaneously from EKG, Journal of the American College of Cardiology: Cardiovascular Imaging, October 13, 2021, DOI: 10.1016 / j.jcmg.2021.08 .004.
About the Mount Sinai Health System
The Mount Sinai Health System is New York City’s largest academic medical system, comprising eight hospitals, a leading medical school, and an extensive network of outpatient practices in the greater New York City area. Mount Sinai advances medicine and health through unparalleled education and translational research and discovery to provide the safest, highest quality, most accessible and equitable care, and the best value of any health care system. health of the country. The health system comprises around 7,300 primary and specialist care physicians; 13 outpatient surgery centers in joint venture; more than 415 outpatient offices in the five boroughs of New York, Westchester, Long Island and Florida; and over 30 affiliated community health centers. Mount Sinai Hospital is ranked on US News and World Reports “Honor Roll” of the 20 best American hospitals and is the first in the country by specialty: n ° 1 in geriatrics and top 20 in cardiology / cardiac surgery, diabetes / endocrinology, gastroenterology / gastrointestinal surgery, neurology / neurosurgery, orthopedics, pneumology / Pulmonary surgery, rehabilitation and urology. New York Eye and Ear Infirmary of Mount Sinai is ranked # 12 in Ophthalmology. Mount Sinai Kravis Children’s Hospital is classified under US News and World Reports “Best Children’s Hospitals” among the best in the country in four of the ten pediatric specialties. Icahn School of Medicine is one of three medical schools that have distinguished themselves by several indicators: ranked in the top 20 by US News and World Reports “Best medical schools”, aligned with a American News and World Report Honor Roll Hospital and No. 14 in the country for funding from the National Institutes of Health. Newsweek “The World’s Best Smart Hospitals” ranks Mount Sinai Hospital # 1 in New York City and in the top five in the world, and Mount Sinai Morningside in the top 20 in the world.
For more information, visit https://www.mountsinai.org or find Mount Sinai on Facebook, Twitter and YouTube.