Straight from the heart

The human body emits millions of physiologic signals, much of which is conventionally believed to be noise, but what if these signals could provide a deeper understanding of our biology and distinguish a healthy heart from a diseased one, even before symptoms appear?

That is precisely what Analytics 4 Life (A4L) is trying to achieve. A4L is using advanced signal processing techniques to identify and assess diseases, particularly coronary artery disease.

Traditionally, patients have to endure a nuclear stress test to evaluate the heart’s blood flow. The invasive test can take hours to days to deliver test results.

A4L’s approach involves machine learning on non-invasive, passively collected signals for heart function. The process of signal collection results in no discomfort to the patient, and is easy to use for the health care provider.

If successful, A4L’s device will streamline the process of identifying cardiac dysfunction to mere minutes and have the results available to the patient’s doctor the same day, with no exposure of the patient to radiation.

A4L is the brainchild of founder and chief scientific officer Sunny Gupta. Coming from a family with a history of heart disease, he wanted to know what information might be taken from the heart’s signals.

The idea derived from the research of Prof. Terence Ozolin at Queen’s University who sought Gupta’s help in identifying diseased hearts in developing mice with pre-existing conditions. Gupta used machine learning to determine which mice would eventually demonstrate the conditions they were born with.

“With 50,000 physiologic signals, we needed more horsepower,” explained Shyam Ramchandani, vice-president, clinical affairs.

“SOSCIP and IBM provided the cloud infrastructure and the computers that we needed,” said Gupta.

The team was able to predict which mice would develop heart conditions later in life.

Gupta founded A4L out of a small room above Chalmers United Church in Kingston, Ontario, and incorporated in July 2012. The company has since expanded to multiple office locations across Ontario and has raised over C$20 million from investors. A4L holds three issued U.S. patents with eighteen more U.S. and international applications pending and have been cleared for testing the device on humans in the U.S. They hope to complete the clinical trials by the end of 2017.

The team perform mathematical extrapolations from the signals and convert them into 3D shapes that have geometric form. They can use these forms to identify and pinpoint disease.

If successful, the device may be able to be used to examine countless other diseases and conditions.

SOURCE: Southern Ontario Smart Computing for Innovation Platform (SOSCIP)