Episode Summary: If there’s any industry ripe for disruption by AI and ML applications, it’s healthcare. This week, we speak with Eleven Two Capital‘s Founder and Managing Partner Shelley Zhuang, whose investment focus (among other spaces) is on innovative healthcare services and applications. In addition to discussing how AI and ML is helping propel genomics, diagnostics, therapeutic treatment, and other innovations into a new paradigm, she touches on what the healthcare space might look like in the next 10 years. For healthcare startups looking to break into the healthcare market, Zhuang doesn’t pretend to have simple answers; however, she identifies commonalities among smart companies that have prepared early for meeting regulatory and other industry considerations. This interview was recorded live in San Francisco at Re-Work’s Machine Intelligence in Autonomous Vehicles Summit in March 2017.
Expertise: Venture capital investment in emerging technology, software engineering
Brief Recognition: Shelley has over ten years of experience in technology as a software engineer, research scientist, business executive, and venture capitalist. Shelley was formerly EVP of Business Development at Ecoplast Technologies, where she oversaw business development & sales efforts in North America. Previously, she was a Principal at DFJ and actively involved in a number of investments, including Ecoplast Technologies, FeedBurner (acquired by Google for $100M), Flurry (acquired by Yahoo for $240M), and several others. Shelley holds a BS in Computer Science and Computer Engineering from the University of Missouri, and a PhD in Computer Science from the University of California, Berkeley.
Current Affiliations: Advisor at Skydeck, ML7 Associate at Creative Destruction Lab, Member Berkeley Angel Network
Big Ideas:
1 – AI & ML Topple Barriers in Bringing Healthcare to the Patient Level
1 – AI & ML Topple Barriers in Bringing Healthcare to the Patient LevelToday, AI and ML platforms applied toward genomics can search a variety of public health databases—ovarian cancer patients, for example—look through the data, find heterogeneous factors, and then sift out subpopulations with the same disease. The goal is then to develop drugs that target each specific population. These and other ML and AI breakthroughs in diagnostics, therapeutics, etc. are opening doors for truly personalized medicine in the decades ahead.
Turning Insight into Action: Companies looking to enter the healthcare space have an invaluable opportunity to create applications and services that personalize routine and preventative care (s of this writing significant percentage of the healthcare AI vendor companies on our platform are focused explicitly on personalized care). The ability to baseline an individual’s various ‘omics’ (biome data, genome data, etc.) and to then coax out risk for disease, shape dietary suggestions (based on “honest” feedback), etc., will give healthcare practitioners a new level of influence in helping shape individual behavior and creating a preventative healthcare system that reduces costs in dollars and human lives.
Interview Highlights:
The following is a condensed version of the full audio interview, which is available in the above links on Emerj’s SoundCloud and iTunes stations.
(3:25) Where where might technologies (like genomics) really have power to improve patient outcomes?
Shelley Zhuang: I can give you a very specific example of a portfolio company called CapellaBio, they’ve essentially built a computational platform that can characterize disease heterogeneity; so imagine ovarian cancer…once they identify this heterogeneity, they’re able to then identify drugs that would be most effective for specific patient sub-populations. So this is really using genomic data or RN-expression data to develop personalized medicines or drugs that can be more effective in a personalized fashion.
(4:45) Articulate what that would mean for the business audience…when you say heterogeneous, speak to what that may mean.
SZ: It simply means if two people have ovarian cancer, on the outside it sounds like the same disease, but if you look at the genomic data or the transcriptome data, which is the expression data of their DNA, it could actually look different, and because they’re different, different drugs might be more effective for one group versus another.
(7:54) Are there also companies leveraging this generic information on the diagnostic side or are you not seeing as much of that?
SZ: Definitely applies to diagnostics as well and — not specifically for our companies — you’ve seen a lot of work in doing blood-based diagnostics, based on the DNA information you can recover from the blood. Recently we’re even seeing saliva, it’s less invasive than a blood sample; so definitely looking at DNA, RNA, proteins, all kinds of biomarker information to diagnose someone as early as possible is definitely a very active area…for example if you look at Grail, the very large spinout from Illumina, they’re very well-funded and they’re trying to use DNA and other information from the blood to be able to detect cancer early.
(11:02) If we look ahead 10 years, maybe a little bit more, where do you suppose the average person…would we bump into the use of genetic information where we don’t today?
SZ: So first of all, it’s not just genomic data, that’s just one source of data, and you’re right—genomic data, the cost of obtaining it has come down drastically…but it’s data like RNA expression data, proteomic data, epigenome data, microbiome data, bacteria in the body, metabolomic data, it’s a wealth of mixed data. As investors, we’re seeing a lot of innovations on the tools side, that would enable us to quickly and cheaply weed out this data, so it’s very exciting for us to see the combination of increasing availability of data married with advances in machine learning and software technologies to extract insights from this data to enable us to solve many different problems in healthcare, that’s diagnostics, therapeutics, treatment and care, clinical workflows…in term of bringing it down to the patient level, we’re already seeing genomics being used for doing pregnancy-related testing…and I think in the future, as I’d mentioned earlier, I’d love to have an annual blood or saliva test that can do cancer detection..if we can achieve that, I think it would be wonderful.
(16:53) How do you see these ‘omics’ folks…selling into the “healthcare” proper, what do you see smart teams are doing to figure out how to get through the maze?
SZ: Maybe we can just talk about diagnostics companies…I think it’s sort of a long process, even though they’re not a pharma company, but they need to first get analytical validity, so basically how does this compare to a gold standard…then you need to get regulatory clearance with the FDA, and then you need to get product market fit, and then lastly you need to get paid coverage reimbursement from the insurance companies, and each of these is a very lengthy topic on its own, but on the product market fit – I think it’s absolutely critical that companies need to clearly understand where their solution fits into existing clinical workflow…and that can only be done through a lot of interviews, a lot of meeting with key opinion leaders, and it would be great if you could get several KOLs (key opinion leaders) to get behind your value proposition; another way we’ve seen companies get product market fit is partner with an existing larger company…and to get their new technology into the market.
Note: If you’re interested in learning more about how to best handle the regulatory and insurance reimbursement side of the healthcare market, listen to the end of the audio.
