AI to Predict Future Cardiac Arrests

Introduction

Have you heard about AI to predict future cardiac arrests?

To improve the precision of forecasting cardiac arrest, researchers from Johns Hopkins University have developed an Artificial Intelligence (AI) system based on the raw images of patients’ hearts and data on their demographics. This technology, which is dependent on raw images of patients’ hearts and patient backgrounds, will revolutionize clinical decision-making. It will also improve survival rates from sudden and deadly cardiac arrhythmias – one of medicine’s most lethal and puzzling conditions.

Sudden cardiac death due to arrhythmia is responsible for as many as 20 percent of all deaths globally. We know little about why it is happening or how to tell who is at risk,” quotes senior author Natalia Trayanova, the Murray B. Sachs Professor of Biomedical Engineering and Medicine.

 

Neural Networks to Build an Exclusive Survival Assessment

The team has been the first to utilize neural networks to build an exclusive survival assessment for every patient with heart disease. These risk measures provide high accuracy of the possibility of sudden cardiac death over 10 years and when it is most likely to happen. The deep learning data science technology is known as Survival Study of Cardiac Arrhythmia Risk (SSCAR). Cardiac scarring caused by heart disease often results in lethal arrhythmias and is the key to the algorithm’s predictions.

There are so many patients who are not at any risk of sudden cardiac death and receive defibrillators that they do not require. And there are high-risk patients that are not receiving the treatment they require and can die in the prime of their life. This algorithm helps determine who is at risk for cardiac failures and when they may occur. The AI to predict future cardiac arrests algorithm thus helps doctors decide what needs to be done exactly.

How does the Technology Work?

The team uses contrast-enhanced cardiac images that envisage the distribution of scars from hundreds of patients at Johns Hopkins Hospital. The technology uses cardiac scarring to train the algorithm to identify patterns not visible to the naked eye. Present clinical cardiac image analysis has provided only simple scar features such as mass and volume.

Recently, physicians from the Society of Nuclear Medicine and Molecular Imaging discovered that analyzing coronary quantitative coronary plaque characteristics and 18F-NaF uptake on PET through an AI model display many signs of heart attack risk.

Another AI approach that a Mayo Clinic research team uses is that of voice biomarkers to describe any potential heart issues. The approach works by giving all patients access to a smartphone app where they enter three 30-second voice recordings. Evaluating elements such as cadence, pitch, and frequency, the app uses the recordings to identify artery clogs.

The algorithms’ predictions have been significantly more accurate on each measure than doctors’ predictions. Also, an independent patient cohort from 60 health centers across the United States confirms the accuracy of these predictions. The platform has the potential to be installed anywhere with different cardiac histories and imaging data.

Conclusion

Artificial Intelligence can help resolve some of the most significant challenges facing healthcare today, such as physician burnout, managing costs, and health equity. AI to predict future cardiac arrests is an example of how helpful AI can be. AI solutions provide healthcare professionals the tools and time they require to offer better care to more patients worldwide.