AliveCor, Mayo Clinic developing AI screening test for heart condition that causes sudden death

By Jonah Comstock
02:00 pm

Smartphone-connected ECG maker AliveCor is teaming up with the Mayo Clinic once again, this time to develop algorithms to screen for Long QT syndrome, a heart condition that can often cause sudden death, especially in children. The partnership also includes an additional, undisclosed investment in AliveCor from Mayo Clinic.

Long QT syndrome is named for the interval between the Q Wave and the T Wave on an ECG.

“For a number of people, they have a genetic mutation that causes the extension of the QT interval and they die,” AliveCor CEO Vic Gundotra told MobiHealthNews. “It’s particularly tragic when this happens to children. You have no warning. What you find out is that their kid suffered from genetic Long QT and it was never caught. Parents spend much of their efforts afterward trying to educate other parents and to put into place screening programs to catch these things. Because if you know you have genetic Long QT, there’s lots of ways to treat it and it’s very very treatable.”

AliveCor is working with Dr. Michael Ackerman, one of the world leaders in Long QT research, to combine AliveCor’s largest ECG dataset and machine learning experience with Ackerman’s existing work on Long QT.

“The dream here is one day we will take the expertise of the Mayo Clinic and the expertise of Dr. Ackerman and dramatically combine them with the AI work we’ve been doing, which would lend to an inexpensive, noninvasive screening test, that one day even a coach who isn’t a doctor, with FDA approval, could inexpensively screen kids and potentially save thousands of lives.”

In addition to genetic Long QT, the syndrome can also be triggered by drug interactions. This means the screening test could also be used by pharmacists before they give patients a drug that could potentially trigger the syndrome. The company hopes to have an FDA-cleared version of the screening test within a year.

In the long term, AliveCor hopes to combine the new algorithm with its existing AFib algorithm, the algorithm the company is developing with Mayo Clinic for high blood potassium (hyperkalemia), and other algorithms.

“We’ll connect all these things,” Gundotra said. “As we improve our algorithms to include more conditions, whether it be hyperkalemia or long QT, it just makes our device more and more valuable and more important for you to have in your home. We believe one day it will be as prevalent as the thermometer. Just like you have a device to check your kid’s temperature at home and know that if it’s dangerously high you should go to the hospital, you similarly will have a device in your home that checks your family members for heart conditions. This is the number one (deadliest) disease in the world, and today people don’t have the tools to check their heart and know what’s really going on.”

The combination of machine learning and small, accurate mobile-enabled ECG readers could lead to a huge paradigm shift in early detection and prevention of heart conditions. A recent study at Stanford University showed that a machine learning algorithm could outshine cardiologists at detecting a number of conditions, given a large enough dataset to train on.

Despite all this, Gundotra says that worries about AIs replacing human doctors are misplaced.

“If we went outside and were walking on a dark path, which one of us would say ‘Hey, I’m not going to use a flashlight because it might replace human vision,' it would be stupid. We would all take the flashlight because it supplements human vision and makes it much easier to walk at night. And the tools we’re developing will be exactly like that. These are not going to replace your cardiologist, but they will allow your cardiologist to see further than they’ve ever seen before and they’re going to become a key way to manage your health and supplement your doctor’s oversight.”


The latest news in digital health delivered daily to your inbox.

Thank you for subscribing!
Error! Something went wrong!