Israeli-based health tech company Cordio has developed machine learning software that can be downloaded to a smartphone and help keeps cardiac patients out of the hospital.
One day in the future, a patient at risk for heart failure may find themselves waking up in the morning, opening up his smartphone and uttering these words: “The cat sat on the ship. David is a big chef. Jeff plays the guitar.”
It’s a simple daily habit that could save their life, because one day after repeating their daily refrain, their doctor might be notified that a patient is at risk of heart failure without immediate care.
That’s the dream of Tamir Tal, the founder of Israeli-based artificial intelligence startup Cordio. Its app, HearO, uses machine learning algorithms to detect changes in a patient’s voice that suggest there’s fluid accumulation happening in the lungs before a patient even notices any symptoms. By catching it early, doctors can intervene with medicine and prevent a patient’s condition from deteriorating to a point where they need to go to the hospital.
“This situation became the Holy Grail of the remote monitoring industry,” says Tal, 50. “To find a way to identify patients when they’re getting close to hospitalization because the patient doesn’t feel it.”
Millions of Americans are at risk for congestive heart failure, a condition in which the heart muscle is too weak to pump all of the blood the body needs. The result is that fluids build up in the extremities and lungs which, if left untreated, can lead to serious complications that require hospitalization. If caught early, patients can avoid a trip to the hospital but it’s often hard to diagnose until the heart failure is serious.
“There certainly are unmet needs for heart failure,” says William Abraham, a cardiologist at Ohio State University’s Wexner Medical Center. “And one of them is to closely monitor our heart failure patients remotely and know when they’re accumulating fluids in their lungs and at risk for hospitalization.”
So far, clinical studies have shown that the app can predict such events in patients about 80% of the time. Those studies, conducted in Israel, have gained the app marketing authorization from regulators in Europe and Israel, the company says. The company says it’s pursuing marketing authorization with the FDA as a breakthrough device after its current U.S.-based clinical trial, which features a larger, more diverse population, is complete.
Abraham, who’s currently conducting the company’s trial, says that if the results bear out and the app goes on the market, it offers a unique solution in the market for heart failure patients. “What’s exciting about this one is its simplicity, its elegance, and that’s it’s ready to be universally deployed to our heart failure patients,” he says. “It doesn’t require any hardware or devices in the home–all you need is a smartphone.”
“Everyone talks about artificial intelligence, machine learning and remote diagnostics and this company had it all.”
Born in Israel, Tal began his career as an attorney focused on financial transactions such as M&A and venture capital investments, and joined Peregrine Ventures as a partner in 2005, a position he still holds. He began serving as COO of Neovasc, which makes medical devices for cardiac patients, in 2007. That position, he says, inspired his fascination with medical devices.
In 2010, left his role at Neovasc when the company was acquired by Medical Venture Corp. In 2017, he joined Cordio as the CEO. The company was founded in 2013 by Aviv Lotan, Sigal Kremer-Tal, Ze’ev Schlik, Chaim Lotan and Shelley Lotan and went through the Incentive Incubator program, which is operated by the Israeli government.
The inspiration for the Cordio app, Tal says, came from cofounder Chaim Lotan, who is a cardiologist. He and his colleagues had noticed that when a patient had too much fluid in their lungs, it was actually audible in their voice. But generally by that point it was too late to prevent a trip to the hospital. So the idea behind Cordio’s app is to see if software could detect those vocal changes earlier.
One of the key features of the machine learning aspect of the app is that the patient is recording a baseline of phrases when they’re known to be clear of fluid, using phrases featuring a wide range of sounds. That baseline makes the changes easier for algorithms to detect, and they’re spoken in the morning because fluid accumulation is easier to detect when a patient has been horizontal for several hours sleeping.
After coming on as CEO, Tal hired Ilan Shallom as the company’s Chief Technology Officer. Shallom has been doing speech recognition research since the 1980s and founded speech recognition company NSC in 1993, which was later acquired by AudioCodes in 2010. “I courted him for a long time,” Tal says. “Longer than my wife–it took about a year.”
So far the company has raised about $25 million in capital from Peregrine Ventures, Ceros Financial Services and other investors. Tal says that the company will likely do another round of fundraising this year to carry it through the FDA approval process and the beginning of commercialization of its product.
For Mark Goldwasser, CEO of Ceros Financial, his company’s investment was its clear vision for how to implement its technology in a way that wasn’t just hype. “Everyone talks about artificial intelligence, machine learning and remote diagnostics and this company had it all,” he says.
The company’s currently finalizing recruitment for its final U.S. clinical trial, which it expects to have at 30 sites around the country, which will give it a chance to see how the app works with a diverse range of patients. Assuming good results from those trials, Tal says that his company is aiming for FDA clearance by the second quarter of 2024.
Once the technology proves out for congestive heart failure, the company intends to expand its voice recognition technology to apply to other diseases. For example, it’s testing predictive software for Covid and COPD patients who might be at risk for hospitalization to catch potential issues before they require a visit. But he sees each as a separate challenge, not something that’s necessarily just another line of code in an app, and is planning his company’s strategy accordingly.
“I think the biggest difference between us and other digital health companies is that we’re thinking like a medical device company,” says Tal. “We started working with physicians early on in the game to develop the product and understand what they need.”
MORE AT FORBES