Despite advances in treatments in recent years heart disease remains the leading cause of death in the US. It accounts for one in three deaths in this country, and many people are not even aware they have a problem until they have a heart attack.
One of the early warning signs of danger is a heart arrhythmia; that’s when electrical signals that control the hearts beating don’t work properly and can result in the heart beating too fast, too slow, or irregularly. However, predicting who is at risk of these arrhythmias is difficult. Now new research may have found a way to change that.
A research team at the Institute of Molecular and Cell Biology in Singapore combined stem cells with machine learning, and developed a way to predict arrhythmias, with a high degree of accuracy.
The team used stem cells to create different batches of cardiomyocytes or heart muscle cells. Some of these batches were healthy heart cells, but some had arrhythmias caused by different problems such as a genetic disorder or drug induced.
They then trained a machine learning program to use videos to scan the 3,000 different groups of cells. By studying the different beating patterns of the cells, and then using the levels of calcium in the cells, the machine was able to predict, with 90 percent accuracy, which cells were most likely to experience arrhythmias.
The researchers say their approach is faster, simpler and more accurate than current methods of trying to predict who is at risk for arrhythmias and could have a big impact on our ability to intervene before the individual suffers a fatal heart attack.
The research was published in the journal Stem Cell Reports.