The expertise of a Mayo Clinic cardiologist who co-invented a new way of gleaning hidden health information from electrocardiograms has been wedded to a Silicon Valley digital health start-up with a device that can generate ECGs for consumers using a smartphone.
 
One goal of the collaboration announced in October between Mayo and AliveCor of Mountain View, California—which includes an equity stake in the company—is to apply proprietary machine learning algorithms developed by Dr. Paul Friedman to ECGs generated by users of AliveCor’s Kardia Mobile device.
 
The Mayo algorithms can analyze the heart signals with a precision unavailable to the naked eye, noting tiny changes, and through that can accurately predict a user’s blood potassium level—all without the need for a blood test, as is now necessary.  That’s important because changing levels of potassium in the blood provide early warning signals for a range of conditions including arrhythmias, heart failure and kidney failure.
 
The ability to non-invasively read those tiny changes quickly could prove to be an invaluable tool in the treatment and prevention of those problems, and, if ultimately successful, be a boost for the digital health market, which has been slow to realize its commercial potential.
 
“We look forward to collaborating with the Mayo Clinic in the interest of giving people a more accurate, holistic view of their internal health,” AliveCor COO Doug Biehn said in a statement issued to TCB. “The Mayo Clinic has invested financially in AliveCor, and the institution is using our technology in its research.”
 
Dr. Friedman, vice chair of cardiovascular medicine and director of the implantable device lab at Mayo Clinic, said the collaboration is an exciting one because AliveCor’s own FDA-cleared mobile technology succeeded in creating reliable ECGs on the fly, thus making it possible for anyone with a smartphone to generate one.
 
“AliveCor’s system is ubiquitous because you can attach it to the back of a phone, and people basically have their phones with them all the time,” he told TCB. “You hold it in your hands for 30 seconds and it records a signal. Imagine how powerful it would be if that signal could give you your blood potassium level.
 
“We’ve demonstrated that (the prediction algorithms) are accurate when tried on small numbers of people who are on kidney dialysis, and now we’re testing it further. The way it works here is that the ECG signal is uploaded from the smartphone to a secure cloud, and then downloaded at the back end, where additional signaling processing can be done on it.”
 
Six-year-old AliveCor is headed by former Google and Microsoft executive Vivek “Vic” Gundotra, well known in Silicon Valley for leading the development of the Google+ social networking platform and Google Maps before that, following 15 years at Microsoft. Its Kardia Mobile device is a low-cost, clinical-grade heart monitor that fits onto the back of smartphone and is also available was a wearable wristband. It employs an ECG device to provide an instant analysis for a range of heart measurements, such as detecting atrial fibrillation and normal sinus rhythms.
 
The company’s venture backers include Khosla Ventures, Burrill & Company, Qualcomm Inc., and the Oklahoma Life Sciences Fund. Mayo has now joined that list of financial stakeholders via the new collaboration, under which it is trading its patented know-how in ECG processing for an equity stake, a Mayo spokeswoman told TCB. It also has a royalty deal in place.
 
“They’ve reliably worked out the kinks in generating clean ECGs on a mobile device,” Friedman continued. “When you’re in a medical environment, things are very controlled and you can get a good signal. But when you’re at home, it’s a lot harder. But the system they’ve developed is very accommodating to the algorithms we’ve developed – we can effective process those signals to get meaningful information.”
 
It has been known for some time that major abnormalities in blood electrolytes such as sodium and potassium and analytes such as glucose result in “deformed” ECG waves, but a system patented by Friedman and three Mayo colleagues in April takes the idea further.
 
It provides for “non-invasively measuring analyte levels” in blood by averaging a series of ECG readings to identify subtle differences – changes that are “clinically important” but not readily noticeable to the naked eye.
 
The technology combines “signal averaging” with an algorithm that can correct the signals to account for such variables as body position changes, respiration, and ectopic (skipped or abnormal) heartbeats, thereby providing sufficient sensitivity to detect small changes of blood electrolytes.