The human heart beats about 100,000 times a day. This constant rhythm is what keeps us alive and enables the functioning of our bodies.
But what happens when your heart stops beating?
Each year, over 350 thousand people in the United States experience unexpected cardiac arrest. Most do not survive until they reach the hospital.
Such deaths are known as sudden cardiac death and are generally defined as sudden and unforeseen death, which occurs within an hour because of heart problems.
According to estimates, sudden cardiac death kills up to 300 thousand Americans each year and is believed to be responsible for up to 15 percent of all deaths worldwide, reports the report. sciencealrt Report signal.
However, a new study published in the science magazine Nature shows that there is a way to predict danger and potentially prevent many of these deaths.
Impplantable defibrillators are devices that can detect and correct dangerous heart rhythm disorders before they cause fatal consequences. The problem is that doctors find it difficult to determine which patients need such equipment.
We have a solution to this problem, says Ziad Obermeyer, physician and artificial intelligence researcher at the University of California, Berkeley.
If you knew you were one of the people who were in danger of dying suddenly, you would go to the cardiologist and place a defibrillator. The problem is that doctors cannot determine who needs it before it is too late, he explains.
In their study, Obermeyer and his colleagues trained an artificial intelligence model based on deep learning to identify the electrical patterns of the heart associated with the risk of sudden cardiac death.
To that end, they used more than 40,000 electrocardiogram recordings ( EKG from Sweden. The artificial intelligence system was also tested with ECG data of approximately 140 thousand patients in the United States, as well as a smaller group of patients in Taiwan.
EKG is a relatively cheap, non-invasive method that records electrical heart activity through electrodes placed in the patient’s skin.
The implementation of deep learning at ECG can offer the benefits of both approaches: a standardized, easily accessible and low-cost census, which can be directly analyzed by an artificial intelligence system to predict danger, without the need for human intervention, researchers write in their work.
The results of the study showed that this approach has considerable potential.
After analyzing a large number of data, the artificial intelligence system identified a high – risk group of patients, which constituted 2.2 percent of the Swedish sample.
Within a year, seven percent of the group suffered sudden cardiac deaths.
By comparison, the traditional method of measuring the fraction of left ventricle injection ( LVEF identified a smaller group of high – risk patients, representing 1.9 percent of the sample. Of this group, 4.6 percent suffered sudden cardiac deaths within a year.
However, only a small portion of patients identified by both methods were compatible. According to researchers, the traditional LVEF could not identify 86.1 percent of high – risk patients who were exposed by artificial intelligence.
These results may be of great importance to preventing death. In high – risk patients identified by artificial intelligence, those with defibrillator implants displayed a rate of sudden cardiac death 54.4 percent lower than expected statistically.
This suggests that equipment can help prevent some deaths. However, researchers point out that the results are observable and do not testify that defibrillators were directly the cause of reduced risk.
Despite that, the findings can pave the way for further research.
In practice, high-risk patients seem to constitute a new, previously unidentified group with which deaths are frequent and potentially preventable, the authors of the study write.
However, scholars also discovered something else.
When using a second model of artificial intelligence to visually present high – risk electrical models, they identified a wave feature that had not been described before and could independently predict the risk of sudden cardiac death.
According to scientists, this feature, called the steep, domestication (slured downstairs), appears in the electrical – wave records of high – risk patients. It can be a sign of close heart problems and can be linked to fibrosis – the creation of internal tissue in the heart muscle, which can prevent the normal conveying of electrical impulses.
Changxin Lai, a biomedical engineer at Johns Hopkins University, who was not part of the study but wrote a comment on it, praised the importance of this discovery.
The fact that an approach based on automatic learning may reveal an unknown pattern in so well – known data proves that deep learning, when combined with other suitable means, can go beyond forecasts and contribute to scientific discoveries, he wrote.
) ECGs have been interpreted by doctors for more than a century, but this study proves that even the most studied data can contain messages that have not yet been revealed.
The signal where the truth speaks before the news!
Source Layer: Signal