Analytics 4 Life Takes The Stress Out Of Coronary Artery Disease Tests

TORONTO, Jun. 8, 2016 /TechPORTFOLIO/ — Data collected and crunched in the cloud aims to replace test that involves radioactive dye injections and running on a treadmill.

For a disease recognized as the most common global cause of death – 8.14 million worldwide in 2013 – coronary artery disease is extremely laborious to diagnose. A Canadian firm, Analytics 4 Life, is looking to cut out a large part of the labor involved – and the danger posed to patients – by leveraging data in new ways.

In a nuclear stress test, patients must exercise on a treadmill to measure blood flow after radioactive dye is injected. Physicians and hospital workers need to spend resources on managing radioactive nuclei. The test can take up to five hours and is only 75% accurate.

Analytics 4 Life, a startup based in Kingston, Ontario, is attempting to develop a much more straightforward method, using machine learning through neural networks and genetic analysis, with physiological information conventionally considered valueless.

“We’ve been able to demonstrate a very simple test that takes about three minutes to do where you don’t have to stress the patients,” says Shyam Ramchandani, Director of Marketing and Business Development at A4L. “It’s just surface electrodes that go on patches on the body: seven of them. Three minutes later, they’re done, and then by the time their patches are off and their shirt’s on the result is on the doctor’s portal.”

This month, A4L is running its first machine learning tests with recruited patients already diagnosed with coronary artery disease, a condition where the vessels that supply oxygenated blood to the heart are obstructed by plaque; if the plaque builds up excessively and hardens, the condition leads to blood clots, angina and cardiac arrest.

After crunching what A4L calls “phase energy” data – which draws on a wide array of physiological signals – from the electrodes, and generating a formula based on the results, the company will then test blind on another selection of patients to see if their formula is an effective predictor.

“‘Phase energy’ is purely a mathematical concept,” explains Ramchandani. “There is no current physiological description of this. We will be the first to demonstrate this.”

The test itself can be administered from a “phase energy signal recorder”, an iPad Mini adapted with proprietary technology to connect to the electrode input. The recorder transfers the data to the cloud. The front-end software and the data processing lives on IBM SoftLayer, and important code infrastructure is hosted on IBM’s Bluemix platform.

This setup makes it all portable. “You technically can take our test anywhere you have a 3G signal,” says Ramchandani. “You wouldn’t have to fly people in from remote areas to a place that has a special camera.” According to Ramchandani, IBM infrastructure is ideal for handling healthcare data. “We have an almost off-the-shelf HIPAA compliant tool. When you’re collecting medical information, you have to either de-identify it in a way that it can’t connect it back to the patient, or it needs to be hosted on and transmitted on infrastructure that’s been validated for security purposes.”

A4L completed series A funding last August for CA$10 million, and is hoping to complete series B at the beginning of 2017 to help it fund commercialization activity and a pivotal clinical trial.

That pivotal clinical trial will support their next key step: the FDA approval process. “If you can’t get through regulatory affairs in an efficient manner and get the kind of reimbursement you need, it’s not going to be a business,” Ramchandani says. The company has made senior level hires in order to facilitate interaction with the FDA.

Another option for A4L is expanding in Europe. (They will not start with Canada initially, because the market is too small.)

Although a price for the test hasn’t been set, A4L says that the overheads for its new system, once approved and functioning, are going to be astronomically less than the status quo. “We have no regulated nuclei that needs to be injected, purchased, or handled,” says Ramchandani. “You don’t need a specialized technician, or a specialized camera.”

And no more running on a treadmill, either.

SOURCE: techPORTFOLIO