Article Date: March 1, 2016
By Ian Coutts
People talk a lot today about “innovation ecosystems,” those nurturing networks that encourage technological breakthroughs and incubate startups. The story of A4L (Analytics 4 Life) and Shyam Ramchandani’s involvement with it, is a virtual case study in how these ecosystems can create opportunities and exciting new products.
Every year in the United States doctors and medical professionals administer ten million stress tests. This aptly named procedure is designed to test patients’ coronary health by measuring their normal heart rate, and then elevating it to see how effectively their heart is working. The process can be particularly hard on those already suffering from coronary problems. Cost too is a major issue, particularly with nuclear stress tests (which account for eight million of the total), where the patient is injected with a radioactive dye, a procedure that requires that a hospital possesses some sort of nuclear facility and an expensive, specialized camera. Stress testing typically costs between $1500 and $2000.
If A4L has its way, however, this sort of sophisticated coronary testing will soon be done for less than half the current cost, using a set up no more complicated than an ECG, just seven sensors placed on the patient like band-aids. And return the results in minutes.
A pharmacologist by training (he holds a PhD from McGill), Ramchandani had worked on a number of biopharma startups in California before arriving in Kingston in 2008 (recently married, his wife was finishing her final year of residency at Queen’s; he decided to take advantage of his time here by doing a one-year MBA at the Smith School of Business).
It was while teaching as an adjunct in the department of ophthalmology in 2011(his wife had taken a post teaching in the same department) that Elspeth Murray, the Associate Dean of MBA Programs at the business school approached him with an interesting opportunity. Two off-campus researchers were trying to develop ideas that one of them, Sunny Gupta, had worked on in the armed forces. Gupta, an RMC graduate, had been stationed at CFB Kingston creating systems that would utilize very faint signals (such as the acoustic track of a submarine) and work out how to interpret them. Now, he was looking for possible civilian applications for his work. Ramchandani had actually rented a lab at Queen’s to pursue possible projects of his own, one of which was supported by the Queen’s Industry Partnerships team, but he began spending more and more time with Gupta and his partner working on what would become A4L; ultimately they moved into an Innovation Park satellite footprint in the Biosciences Complex on the Queen’s campus.
In a pair of coincidences that might strain credulity in a film or a novel, two other pieces of the A4L story soon fell into place. Terence Ozolinš, a Queen’s professor and a friend of Ramchandani’s from his McGill days, approached him with a particular challenge: he and his students had been doing work on a drug known to cause congenital heart problems. They had collected heart function data on rats, but could not find any discernible patterns that alerted them to which rats had a potential problem until they stressed the hearts – this could be bad for a patient and was there a way to identify the damaged hearts without using a stress test? As it happened, Ramchandani had recently been introduced to the Southern Ontario Smart Computing Innovation Platform (SOSCIP) through Queen’s Industry Partnerships with information on how to use the consortium’s computing power. Thinking of the work that Gupta was doing, the data his old friend had, and the offer of computing power, Ramchandani told Ozolinš “You give me these undistinguishable subjects and we can distinguish them for you.” Picking up the gauntlet, Ozolinš and Ramchandani collaborated on two successful industrial research grants. The first gave unlimited access to SOSCIP, Canada’s fastest supercomputer, the other funded the in-life experiments through a program called Ontario Centres of Excellence (OCE) Medical Sciences Proof-of-Principle (MSc PoP).
“This is what an innovation ecosystem is all about,” says Steven Liss, Queen’s V-P of Research. “You get something interesting happening over here and see an interesting possibility over there. The ecosystem brought together talent and expertise, from A4L, the Smith School of Business, the School of Medicine and Industry Partnerships, and resources from Innovation Park, SOSCIP and OCE so that new exciting things could happen.”
“We extracted mathematical features from Ozolinš’ data set comprising 50,000 signals for each of 17 subjects totaling 850,000 signals,” says Ramchandani. “We took 15 percent of the signals. We machine learned on these to create an algorithm, which we then applied to the remaining 85 percent.” They discovered they could predict which animals had the defect with 75 per cent accuracy. Having SOSCIP’s resources really helped them “thoroughly interrogate” Ozolinš’ data, says Ramchandani.
Now their thinking is, they can do the same for humans with coronary problems, using the data that could be collected relatively innocuously and inexpensively. Their initial idea was to create just the algorithm, but when they brought in successful entrepreneur Don Crawford to serve as CEO, he convinced them that manufacturing their own device was the way to go.
In January 2016, A4L announced that it had raised $10,000,000 Cdn as part of its first round of financing. The money will help the firm enroll patients and begin developing its new system starting this spring, at the company’s facility in Raleigh, North Carolina (manufacturing takes place there to achieve FDA approval). “We expect,” says Ramchandani, “to have our algorithm developed by fall 2016.” The firm employs five people there, as well as 20 or so people at three different consulting firms, and has further seven on the payroll, working since June 2015 in Toronto. IBM, one of the founding members of the SOSCIP consortium, is very interested in their work, as is SOSCIP itself, which plans to sponsor three post-doctoral fellowships at A4L this spring.
Ramchandani, who is now also an adjunct in Smith School of Business, is currently dividing his time between Kingston and Toronto. He sees plenty of opportunities in and around Queen’s but for the moment he wants to concentrate on A4L. What A4L is doing with coronary testing could be just the beginning, says Ramchandani. “I think we are on to something big here. The machine learning platform we have created can generate findings from any good data. We are going to create a bunch of opportunities – in cardiology, neurology, cancer -- anywhere signals are collected. They could be echocardiograms, they could be MRIs. If we have good data, that is it’s been collected properly, there are no limits on what it can tell us.”