With a Patient's Virtual Heart, Doctors Predict Cardiac Arrest

null
Image: Hermenegild Arevalo and Natalia Trayanova
From the patient’s scan (blue), a virtual heart is constructed and the presence of arrhythmia, indicated by irregular electrical activation (red-yellow), is revealed.

Sudden cardiac death is a terrifying possibility for patients who have been diagnosed with arrhythmia (abnormal beating of the heart), and the risk is particularly acute after a heart attack. To guard against this outcome, cardiologists implant defibrillators in the chests of all patients deemed at risk, but that leaves a lot of room for error. Identifying those patients who really need the devices is tough.

img
Image: Hermenegild Arevalo and Natalia Trayanova
Arrhythmia risk prediction for two patients with previous injury. In the high-risk patient, the electrical wave is “stuck” rotating around the injury.

Now, a new computer modeling system allows doctors to analyze the unique characteristics of each individual patient and predict his or her level of risk. The new technique out-performed current evaluation methods, and offers a promising example of the “precision medicine” that’s all the rage these days.

The lead researcher in this study is Natalia Trayanova, who described similar work in a feature article for IEEE Spectrum in 2014: “Custom Cardiology: A Virtual Heart for Every Patient.”

The heart is an electrical object: Each heartbeat is triggered by an electrical impulse that travels through the cardiac tissue and causes the chambers to contract and pump blood through the body. But after a heart attack, scar tissue on the heart can interfere with the propagation of those electrical impulses, causing arrhythmias and possibly cardiac arrest. So a post-heart-attack patient may have a defibrillator implanted that continuously monitors the electrical activity of the heart, and shocks it back into normal rhythm if the machine detects an irregularity. 

In a paper published this week in Nature Communications, a team of biomedical engineers and cardiologists from Johns Hopkins University described a study of 41 patients who had heart attacks and received implanted defibrillators. Their risk of arrhythmia had been judged in the typical way: Doctors estimated the “left ventricular ejection fraction,” a measurement of how much blood the left ventricle can pump out of the heart to the body. If that figure is under 35 percent, patients typically receive an implanted defibrillator.

However, not all people with low ejection fractions will go on to develop arrhythmias, and implanting a defibrillator carries risks of its own: There are dangers associated with surgery, possible infections, and the very real possibility that the implanted device may malfunction and shock the person needlessly.

img
Image: Hermenegild Arevalo and Natalia Trayanova

To improve this situation, Trayanova and her colleagues constructed 3-D computer models of each individual patient’s heart based on MRI scans. The models used the particular geometry of each patient’s heart and the location of the damaged tissue caused by the heart attack, and also incorporated information about the electrical conductivity of that tissue. Then the modelers could experiment by stimulating different places on the heart and watching how the electrical signal rippled through the tissue, looking to see if it caused an arrhythmia. 

This was a just proof-of-concept study: All 41 patients had already had the defibrillators implanted long before the models were made, and the researchers followed them for years to determine their outcomes. The computer models were found to be more predictive of arrhythmias than other clinical metrics. Those alternatives include not only the ventricular ejection fraction, but also information gathered through an invasive procedure called an electrophysiological study, in which a catheter is snaked through a patient’s vein to reach the heart and measure its electrical activity. 

With such a positive outcome, though, the researchers hope that computer models will soon be used routinely to assess cardiac patients’ risk of arrhythmias, potentially preventing unnecessary defibrillator implantations. “We believe that computer modeling is poised to transform areas of medicine,” the researchers write in their paper, “and serve as a vehicle to advance personalized approaches to human health.” 

Advertisement

The Human OS

IEEE Spectrum’s biomedical blog, featuring the wearable sensors, big data analytics, and implanted devices that enable new ventures in personalized medicine.

 
Senior Associate Editor
Eliza Strickland
Contributor
Emily Waltz
 
Contributor
Megan Scudellari
 

Newsletter Sign Up

Sign up for The Human OS newsletter and get biweekly news about how technology is making healthcare smarter.

Advertisement