According to researchers, now Artificial Intelligence can identify patients who can die because of a medical issue within a year, after studying their standard ECG tests.
Researchers from Geisinger Health System in Pennsylvania studied the results of 1.77 million ECGs and related records from almost 400,000 patients. The data was then used to compare machine learning-based models that either directly analyzed the raw ECG signals or relied on aggregated human-derived measures (standard ECG features typically recorded by a cardiologist) and commonly diagnosed disease patterns, as per IANS.
The neural network model that was used to analyze the ECG signals was found to be surprisingly far superior in predicting the one-year death risk.
As part of the study, three cardiologists reviewed the ECGs that first appeared normal, and failed to recognize the risk patterns that the AI was able to detect, researchers said.
Brandon Fornwalt, chair of the Department of Imaging Science and Innovation at Geisinger in Danville, Pennsylvania wrote, “This is the most important finding of this study. This could completely alter the way we interpret ECGs in the future.”
There was another study conducted by the same researchers who found that AI-based models can better analyse ECG test results and pinpoint patients who are at a higher risk of developing dangerous irregular heartbeat (arrhythmia). More than 2 million ECG results were studied for the study.
While the vast Geisinger database is a key strength of both studies, the findings should be tested at sites outside of Geisinger, the researchers noted.
Both studies are among the first to use AI to predict future events from an ECG rather than to detect current health problems. This discovery shows that we are on the verge of a revolution in medicine where computers can now empower physicians to improve patient care.