Bay Area-based HeartFlow Is Using 3d Modeling to Produce a Better Noninvasive Test for Heart Disease
Technology with heart.
More than 50,000 patients in, CEO Charles Taylor says they’re just getting started.
To take on America’s leading killer, heart disease, you might imagine our country would stop at nothing to ensure the latest and best technologies are being used to diagnose and treat patients.
According to Charles Taylor, you’d be wrong.
The esteemed Bay Area bioengineer has earned a global reputation pioneering the use of computer simulation methods to analyze medical data, specifically data pertaining to blood flow in and around the heart. “When I started this work, I was really surprised by how poorly existing [noninvasive] tests for heart disease worked,” he recalls, singling out treadmill electrocardiogram (ECG), echocardiography, and SPECT nuclear imaging scans. “The tests have been demonstrated to have false positives and false negatives; it kind of misses in both ways.”
Taylor was speaking to Gentry in January at the Redwood City headquarters of HeartFlow, the health-tech startup he co-founded in 2007. The Stanford PhD and former professor’s own office there sits at the end of a hallway lined with a selection of some of HeartFlow’s hundreds of patents and scientific publications. The company uses software trained with a deep-learning algorithm to analyze a patient’s coronary CT scan (a standard, noninvasive imaging test) and create a 3D model of the coronary arteries to help clinicians determine whether sufficient blood is reaching the heart.
HeartFlow has raised about $530 million to date, and its technology is currently being used in nearly 200 hospitals around the country, as well as more than 50 of the National Health Service hospitals in England and an increasing number of hospitals in Japan—with plans to keep expanding around the world. “Our mission is to give the best noninvasive diagnostic information so clinicians can make the best possible diagnosis for their patients,” Taylor explains.
In essence, Taylor has been working for decades to get to this point. Unfortunately, the tale of his success begins with a near-death experience.
“When I was about 15, I had a ruptured appendix and very nearly died,” he recalls. After several intense weeks in the hospital and feeling inspired by the surgeon who saved his life, Taylor decided he wanted to become a doctor. Over the next few years, however, his interests shifted toward engineering with a focus on “building computer models for physical systems,” he says. Originally, that meant engineering work for GE, where he spent four years before beginning his doctoral studies at Stanford in 1992.
“A little over a year into the program, I saw a flier on campus, a talk on ‘Blood Flow and Your Health’ that was going to be given the same day,” Taylor shares. The speaker was Christopher Zarins, MD, at the time Stanford’s new chief of vascular surgery. “He talked about how much engineers could contribute to [the field of] vascular disease, and that singular event…I thought, ‘This is my calling,’” Taylor says. “I went to see Chris the next day and asked him, ‘Is there a role for me to work on this with you?’”
Zarins ultimately became Taylor’s PhD co-supervisor, and his doctoral thesis focused on using computers to simulate blood flow in arteries. “One of the big ideas I had was to use computer images to create a patient-specific model,” Taylor says. “At that time, in 1995, computers weren’t what they are today, medical imaging wasn’t what it was. The technology wasn’t ready for primetime in 1995, but Stanford was a great place to incubate this idea.”
Taylor stayed on at Stanford for 14 years as a member of the faculty, primarily in bioengineering and surgery. All the while, he continued to pursue the development of computer modeling and imaging techniques for cardiovascular disease research, device design, and surgery planning. “I started thinking, what am I going to do with the rest of my life?” he recalls of a yearlong sabbatical in 2006. “It was unsatisfying to me not to see this technology go into clinical practice and be put to work for patients.”
Determined to change that, Taylor and Zarins co-founded Cardiovascular Simulation, Inc., in 2007 (its named changed to HeartFlow in 2010), the first company-in-residence at the Mountain View-based Fogarty Institute for Innovation, a medical technology incubator.
Thirteen years later, HeartFlow is firmly on the map, and expanding rapidly. In the world of federally regulated healthcare innovations, that’s basically an overnight success story. Taylor explains, “You have to do the science, build the product, go through the regulatory processes, get physicians to change deeply ingrained habit, talk to healthcare systems, and get paid. Nevertheless, we’ve managed to become a software company that’s very agile.”
HeartFlow analysis has been performed on more than 50,000 patients to date, with an average processing time of 2 – 5 hours, depending on the urgency of the case. “Of all the patients that come into it, roughly 50 – 60% of them can be diverted to treatment based on medical therapy; it really reduces the number of invasive diagnostic tests,” Taylor says, citing results of the PLATFORM trial published in 2015 in the Journal of the American College of Cardiology. “What’s more, in the U.S. healthcare system, you’re saving about $3,000 a patient.”
“This is a problem that has massive economic potential,” he continues. “We already spend hundreds of billions of dollars on healthcare, and the youngest of the baby boomers will reach Medicare age in about nine years. We cannot afford to use healthcare services that don’t work, that don’t provide value.”
Looking ahead to the coming year and beyond, Taylor notes the potential applications of HeartFlow’s technology elsewhere in the body, sharing, “We’ve contemplated using this in just about everything.” But he stresses the need for now to stay focused—especially when the focus is on fighting the leading global killer. “It’s enough for now,” he says.
“When I worked at GE, the idea was, if it doesn’t work the way you want it to, you redesign it,” Taylor reflects. “Well, why couldn’t you do that in medicine? I wrote a paper over 20 years ago called ‘Predictive Medicine: Computational Techniques in Therapeutic Decision-making’ that explored that question.”
Before continuing, his gaze extends past his glass-walled office, perhaps over to that lengthy corridor of patents and studies. “It’s hard to describe the feeling now that something you’ve been working on your whole life is having this kind of impact.”