Leo Grady, CEO and Founder of Stealth Venture, on the Applications of AI in Healthcare
Looking for more life sciences stories? Check us out at Spotify, Apple, Soundcloud, or LinkedIn
For the second episode of the season, we spoke with Leo Grady, Ph.D.
Leo has 20 years of experience leading development and commercialization of advanced machine learning, AI, computer vision, diagnostic and digital health technologies that have made a significant impact on healthcare practice. During his tenure as CEO of Paige. AI , Leo led the company to become an industry leader, internationally launched several groundbreaking software products and became the first-ever company to receive FDA approval for an AI product in pathology. Leo is currently CEO in Residence with Breyer Capital and Founder and CEO with Jona.
Prior to joining Paige, Leo was the SVP of Engineering for HeartFlow, where he led full stack technology and product development efforts for HeartFlow’s cardiovascular diagnostic and treatment planning software while also driving HeartFlow’s IP portfolio. Before HeartFlow, Leo served in various technology and leadership roles at Siemens healthcare. Leo is internationally recognized as a technology leader in AI for healthcare, authoring two books, over 100 peer-reviewed scientific papers and having over 300 patents worldwide. Winner of the 2012 Edison Patent Award, he was inducted as a Fellow in the American Institute for Medical and Biological Engineering. Leo received the B.Sc. degree in Electrical Engineering at the University of Vermont and a Ph.D. in Cognitive and Neural Systems from Boston University.
In our conversation, Leo and I chatted about:
- His career journey and starting out in the world of AI
- Transitioning from a larger multi-national company to smaller startups
- The applications of AI in healthcare in cardiology and oncology
- Data privacy and interacting with the FDA for AI solutions
- What company creation and building looks like for different levels of companies and highlights of his new stealth venture
1:40 to 4:40 From the Apple 2.0 to pursuing a Ph.D. in AI, cognitive and neural systems
- On his path to AI, computer vision, and advanced technology: Leo’s father was a virologist leading the New York State Health Department for many years. His interest in computers transitioned to his children we he brought home computers for Leo to play with.
- Leo studied electrical engineering and anthropology in college but did not know how they connected. After further reading into science fiction such as Neuromancer and learning about early thoughts around AI, Leo became excited about its potential and decided to pursue his Ph.D. in AI, neural and cognitive systems
4:41 to 8:36 Early Career at Siemens
- On joining Siemens: After completing his Ph.D. Leo joined the AI group of Siemens Healthcare, a German medical device company that was focused on the software applications for the medical imaging scanners such as CT and MRI. He was impressed by all of the work that was coming out of that AI group and wanted to be where the action was.
- Leo was put on the management track after working in the AI group to develop management, leadership, and interview skills that set him up for leadership positions throughout his career
“What I learned at Siemens was that it is not just the technology that is going to make an impact on healthcare. It was also understanding everything around the technology, patient workflow, user experience, how you connect it to the data sources and more.”
8:37 to 17:00 Diagnosing Heart Disease and HeartFlow’s transformative AI Technology
- On diagnosing coronary artery disease: Heart disease is one of the leading causes of death worldwide. Before a heart attack many patients may have symptoms beforehand such as chest pain, shortness of breath etc. The doctor needs to determine is this actually a heart attack or another medical explanation and if it is a coronary artery blockage, is it severe enough to intervene with a stent or stick to medical management.
- The standard of care before and after HeartFlow: Fractional flow reserve is a procedure done in a catherization that includes inserting a pressure wire into the patient’s artery to measure the pressure difference before and after the blockage. Prior to a catherization an EKG and stress test is also conducted to pick up if the heart is already died. The cardiac guidelines this past year changed to include a cardiac CT because you can visualize the coronary arteries better and see calcification. HeartFlow is a AI tool for the cardiac CT to determine if the calcification is actually coronary artery disease with extreme precision and if the patient needs to have a coronary intervention.
17:01 to 20:15 Building AI products, compliance, and dealing with patient data
- On dealing with patient data: Leo highlight the importance of maintaining privacy and security for patients’ data. He discusses different scenarios in which entities patients interact with can use their data based on the agreements they sign
“As a society we have two competing needs. Once is the need for patient’s to own their data and maintain their privacy. We also have a second societal need which is the ability to develop new technologies to benefit patients, hospital doctors. Aggregation of anonymized data with the right security protocols and consents does give us the ability to do both.”
20:15 to 28:07 Innovating Oncology Care with AI
On the origin of a company that spun out of an academic institution: Leo touches on how Memorial Sloan Kettering got ahead on the rest of the world in digital pathology over five years ago, building a center to start digitizing pathology slides and building AI applications for cancer diagnosis
What is pathology? Pathology is branch of medicine where you take tissue out of the patient, cutting it, mounting it on a slide, staining it, and looking at it under a microscope to make a diagnosis. Pathology is usually the final line of diagnosis for cancer, it is very hard to appeal that.
On being a first-time CEO of a smaller company: Leo joined the company when it had about 15 people and grew it to over 200 people and raised $200 million in funding. He discusses how the CEO role is a reflection of the stage of the company and the needs of the company at a time. He shares how one of the ways it changed was the introduction of new functions: marketing, sales, customer support and more. The other change was specialization that evolved over time with the varying technical teams. As CEO he transitioned from being hands on to bringing in more specialized expertise and delegate work. As the company grew, Leo became more in his comfort zone because it reminded him of working at Siemens.
28:08 to 31:42 AI Technologies Through the FDA Process
On proving safe and effective of AI products: Leo speaks highly of the FDA and their willingness to get innovative products to society. What becomes really important is understanding what the technology does and what it will be marketed to do. To dive into if the AI solution is safe and effective, the FDA will dig into how the company designs the study, if the algorithm is agnostic to the type of machine or scanner, and if the patient population used was representative enough to be safe for everyone in the future.
On interacting with the FDA: HeartFlow and Paige.AI went through the de novo process, classified as class II devices but there was no predicate on the market. Thus, they could not use historical study design to validate. Leo mentions how they received breakthrough designation and that essentially put them on the speed dial with the FDA where you can work with them in a much more interactive way to achieve the first approval.
“The FDA also wants to get good technology on the market. When Paige.AI got FDA approval, the FDA commissioner tweeted about how exciting this technology was to be on the market. At the same time the FDA needs to be convinced that it is safe and effective for the broad U.S. population.”
31:43 to 38:59 The New Stealth Venture and Future of AI Applications
“I think there are a lot of ways in which AI is being developed and used that is not about automation at all but about enabling doctors and patients to do things that they couldn’t do before.”
On his new venture: Leo discusses how his new venture is focused on precision medicine leveraging one of the omics that he believes is underutilized and underpowered. The company will be providing diagnostic information to doctors and patients and be help develop therapies in the future for a wide range of diseases.
On exploring a career in health tech and being an entrepreneur in the space: Leo advises to join a start-up, the earlier the better because it allows you to understand what it is like to build technology. He recommends doing it with people that have done it before and against cautions on starting your own startup
“I have never thought about my career as wanting to be a CEO. I have always looked at as how do we build something to have a transformative impact on healthcare. If I can serve that mission in one way or another, I will play that role. It is really about bringing this technology to the world. There is a lot of work to do.”
