Table of Contents
Could you start by giving us a brief overview of your career journey and what led you to found BillionToOne?
In terms of career journey, BillionToOne is my first, and probably the last job that I will ever have.
I have a very interdisciplinary background. I did my undergrad at Princeton, where I met my co-founder, David Tsao.
We went through this very rigorous program called Integrated Science that approaches biological problems from first principles. In this program, you learn physics, chemistry, computer science, and mathematics; and you use those disciplines to solve challenging problems in biology. It was a very formative experience for us both, because it developed this sense of approaching biology from first principles. We thought that most biology would be like that, but it turns out that it is more of a bubble. Coming out of Princeton, I was as much of a physicist as a biologist.
Based on this educational background, David and I developed a very different, truly interdisciplinary approach towards biology. When I went to Stanford for my PhD, I sought out professors who were working at the intersection of different disciplines. I did my PhD in systems biology using mathematical principles, computer simulations, and single cell time lapse fluorescence microscopy to better understand how cells make decisions.
It's a really fascinating area. You're trying to figure out how cells integrate different signals that are coming from outside. How do they measure their own size? They don't have a ruler. How do they know when it is time to divide? Fundamentally, those considerations are what causes cancer. Because when cells make the wrong decision, they can divide uncontrollably, which is the definition of cancer.
During and after my PhD, Steve Quake was one of the pioneers of the cell free DNA revolution and was looking at this biomarker from blood samples, to understand, in prenatal testing, whether the fetus may have a genetic condition, and then, applications to cancer and other areas. We wanted to understand that particular cell free DNA problem, because we realized that the potential of cell free DNA is so much greater, if we can get to a single molecule level of sensitivity with it.
Both David and I have a very physics-oriented approach towards biology, which really sets us apart from other diagnostics companies. We tend to not use the ‘trial and error’ approaches that are quite common in biology. Rather, we ask what’s the noise that is being introduced by each lab process? How can we reduce that to get down to single molecule level sensitivity?
Chromosome abnormalities were already being detected by other companies, like Natera and others. We also realized that single gene conditions are among the biggest global problems, in particular sickle cell disease and beta thalassemias, which are some of the most common genetic disorders in the world. The World Health Organization puts out regular statements about the global burden of hemoglobinopathies.
We realized that by approaching this from a physics perspective, there may be just enough signal to noise in cell free DNA to detect a single base pair for a single gene condition. With that understanding, you can start developing the technology to remove the noise and get down to that physical limitation of what is possible, so that we can detect those conditions.
Once you can detect them, the applications are quite limitless, because you then improve the sensitivity significantly.
One of the key focuses for the company is early cancer detection. Could you shed some light on exactly how this will be clinically useful and the technical challenges you're working to solve to make that a reality?
Early detection is something that is important because the ideal scenario is that we can catch cancer early, before it spreads. At that point, you can perform curative surgery and prevent cancer to begin with. That is why there's so much interest in that area.
We are approaching cancer diagnostics in a systematic way. Today, we’re not working on early detection. The reason is that we want to get there in a step-by-step manner so that, by the time we are working on that problem, our system is robust from a technical point of view. That is what makes our approach different from other companies in the space.
We started in late-stage cancers, where our sensitivity allows patients to get better targeted therapies and allows medical professionals to track their patients’ tumor burdens much better than with scans alone. Now, we are working on bringing that to the MRD setting (minimal residual disease), where being able to detect, post-surgery, whether a minute amount of DNA remains, which is otherwise undetectable to scans. That will allow doctors to know whether the patient needs adjuvant therapy, chemotherapy, or immunotherapy.
But the technical problem that we are solving for MRD today is very similar to an early detection problem. In both, you are trying to detect a very small amount of tumor DNA, and you are trying to detect it among many other pieces of DNA that are coming from other tissues.
Our path is going from late stage to MRD, and then to early detection, because we believe that once we can solve the MRD at an ultra-sensitive level - sub hundred part per million - we will have solved a technical problem of early detection. It will still require years of costly clinical studies, in the region of a hundred to two hundred million dollars for early detection. However, we want to do that once we have proven that the technical part is resolved.
It’s a very different approach to many others where companies either go into these areas in parallel, or they upgrade their assay after running expensive clinical trials. Our approach is such that when we decide to run the clinical study, we will know what the performance is going to be.
That has been our approach throughout our past. Five years ago, when we launched our prenatal test, we mathematically predicted what the performance was going to be in our clinical studies, and the results ended up being exactly what we predicted they would be.
Could you give us an idea of how a newborn, who goes through one of these maternal blood tests, has their life improved from an earlier diagnosis?
If you detect these conditions prenatally, you can act either in utero or immediately after the baby is born. We are seeing more case studies where the baby has cystic fibrosis that is detected prenatally, and the mother is put on CFTR modulators. And not only is the baby no longer going to the NICU, sometimes the baby is even passing newborn screening, which is remarkable compared to a typical cystic fibrosis baby.
For SMA, after birth, every single day, there is irreversible damage that happens in the newborn. There is a difference between starting the therapy at six weeks of life versus as soon as the baby is born. And that difference is so critical in ensuring the best outcome for the baby
Importantly, this is also an expensive, $2.1 million therapy, so it is important to start the insurance approval process during pregnancy. Trying to get approval at six weeks of life with a newborn can be extremely challenging. Having a two-and-a-half-year-old, I can only empathize with the challenge. Those first few weeks are a blur. I can't imagine dealing with an SMA diagnosis during that time.
And as I mentioned, even if you do it as early as six weeks, the big problem is that, with every passing day, there is irreversible damage. Early diagnosis can significantly improve the outcomes for these babies' lives.
Could you expand on the importance of quantifying single DNA letters from a single molecule? How does that level of precision shape your roadmap?
Let me first talk about cell free DNA which is truly a remarkable biomarker and microcosm of the entire body. You get DNA from every single tissue and the epigenetic markers that show you how that tissue is behaving. But it is also incredibly rare.
As you move from detecting large genetic changes, like chromosome abnormalities, toward much smaller changes, like single gene variants in prenatal testing - and similarly, as you move in cancer from late-stage disease to early-stage disease and eventually early detection - the amount of tumor or fetal DNA in the blood becomes extremely rare. It becomes a true needle in a haystack problem.
As the initial starting amount of material is so low in quantity, every molecular diagnostic that works with cell free DNA must greatly amplify that DNA before you can sequence it and see what you have.
The primary issue is that amplification is required in any cell free DNA methodology, no matter which company is doing it. But that amplification process uses enzymes found in nature, like polymerases, and polymerases are inherently error prone. If they were not error prone, we would never get mutations or cancer to begin with.
When you amplify the DNA, you introduce errors during that process. When you analyze it, you end up having this fundamental problem where these mutations have been introduced by your methodology. How do you differentiate those errors from the original mutations in the sample?
This is indistinguishable from the real mutations, and that sets a fundamental threshold on the sensitivity that any molecular diagnostic using cell free DNA can achieve.
If you look at early detection or MRD products that are tissue naive, the performance across the field is very similar. Early-stage detection rates for many cancer types are often around 50 percent sensitivity, even though different companies use very different methods. Some use fragment patterns, others use methylation, others count sequencing reads, and they look at very different numbers of genomic locations. But they still all reach roughly the same sensitivity, which is not as high as anyone wants.
The reason is that a fundamental threshold is being introduced by the lab processes, specifically the amplification that has to occur before analysis. That is where our technology comes in. We add single molecule controls that let us see where errors and amplification biases are happening. Then we can correct them in the sequencing data and remove as much noise as possible.
The main difference is that in cell free DNA we are not limited by signal because you can always look at more locations. With sequencing costs coming down, companies are increasing the number of loci they measure in hopes of getting to a 90 percent sensitivity for early cancer, but the problem is not signal, it is noise. That is where our technology shines, because by reducing noise we improve the signal to noise at every genomic location we measure.
How dependent are you on Illumina machines?
Our technology is sequencing platform agnostic. From that perspective, many of our assays have been validated across more than one platform, which is very important because the Illumina instruments are great, but every sequencing platform provider runs into consumable problems from time to time.
When you are at single molecule level sensitivity, you are going to see those error rates and problems show up in your data. We are usually the ones who go to these sequencing companies to tell them that there is something wrong in their reagent manufacturing. They say there is nothing wrong. We say that there is something wrong, and we can show how clear it is, thanks to our data.
Having more than one platform as a clinical lab really allows you to be flexible by being validated across more than one platform.
Is your current lab enough to serve the US for prenatal tests?
In our current setup, we have one prenatal lab and one oncology lab, and we can grow another 3x or so before we run out of space. We are currently building a lab that is going to be three times larger than the largest one that we currently have.
That will be sufficient for close to 8 million tests a year. I estimate that it is close to 50 percent plus market for the prenatal and oncology testing, with the exception of cancer screening. Once you get to early detection though, even 8 million tests a year is not going to be enough, because you are looking at population level testing at that point. But until then, which can be many years from now, the lab that we announced in Austin, where we had the ground-breaking, which will be functional by the end of 2027, should allow us to continue to grow.
How much resistance have you had from clinicians based on cost?
This is important. When we design an assay, we prioritize making cost a non-issue for patients. We accept all insurances. Even if the insurance does not pay, we still work with every patient to make it accessible.
But that also requires one to have these novel tests to be designed and launched one at a time so that the previous test already has broad coverage and reimbursement, allowing us to market the more novel test that we have developed.
For us, we have always worked with every patient to make our tests accessible and affordable regardless of their insurance and their financial situation. That has been one of the important drivers of our growth, so that the physician is making this decision based on clinical factors, and not on cost alone.
Even then, when we were early on, being out of network was a big headwind. Even when a test is made accessible and available to everyone, a lot of health systems do not want to use out-of-network labs. Moreover, a laboratory usually gets paid less, sometimes only 15-20% of Medicare/Medicaid prices, when it is out-of-network with commercial insurance.
Now we are processing 600,000 plus tests per year. As we’ve increased that test volume significantly over time, we have been able to become more and more in-network, because one driver to convince an insurance company to take you in-network is test volume, as it proves utilization by physicians.
As we have become more in-network, it is easier for physicians to say that they are going to use BillionToOne tests. And that is driving even higher test volume, which is driving even more in-network contracting with different insurance companies.
Today, as we shared in our earnings callwe are at 235 million contracted lives in the United States, and that continues to increase every quarter as our test volume increases.
Today, not only is cost no longer an issue, but we are also 70 percent in-network with insurance companies, which allows us to work with health systems that really prefer in-network labs.
Between reproductive health and oncology and other therapeutic areas, how do you prioritize which market segments to pursue?
For us, it really is about doing things step-by-step. In prenatal, we did not start with NIPT for Down Syndrome and other aneuploidies, similar to all other NIPT companies. We started with single gene conditions because no one else was doing it. It was a difficult but important problem, and we knew we could solve it. From there, we added NIPT because physicians wanted a single lab, and a single blood draw that could cover all the recommended genetic conditions. Now, we’re expanding into areas where our technology can give them even more information.
In oncology, we took the same approach. We started with late-stage cancers, we are going into MRD next year, and the longer-term goal is early detection. The idea is that every product de-risks the next one. If you can do single gene testing, doing NIPT is relatively easy. If you are doing late-stage cancer testing, you learn a lot about how early-stage cancers behave, because you still see some of those patients, and once you launch that product, you already know what performance looks like when you eventually move into early detection.
When we choose big areas of commercialization, the technology could apply to many other fields, but each new area would require building a new commercial team, and having a great product is maybe 1 percent of the problem. The other 99 percent is getting that product to standard of care, and we are still working on that even after five years.
On the prenatal side, we have grown to be the second largest prenatal lab in the country and are growing fast, but, today, many of our tests are still not the standard of care everywhere.
Healthcare takes 10 years to change. Even with the best products, you need huge investment in commercialization, clinical studies, reimbursement, and market access. Where prior products are effectively enabling the next ones, we move step by step. That works in two ways. It technically de-risks the next product, and it lets you launch it in a fiscally responsible way, because your earlier products are already profitable. You make the first product profitable, then the second, then the third, and so on, so that earlier products fund the launch of something that is not reimbursed yet.
This is one of the biggest problems in molecular diagnostics. At one point, the Gates Foundation sponsored a study that looked at why there is limited diagnostics innovation, as it's so critical to personalized medicine. If we increase the investment in diagnostics, a lot of healthcare problems would get resolved, but there are not, as of yet, many large and profitable companies in molecular diagnostics.
I think it comes down to two things. First, if you do not have technical and clinical differentiation, it is very hard to become profitable because the market becomes highly competitive. That is why we focus on products where we know we have technical and clinical advantages. Second, you need that step by step, de-risked expansion, so that each previous product helps grow the commercialization, market access, and reimbursement of the next one, allowing you to stay profitable at every stage.
Could you give us an update on your progress with the FDA and the challenges and milestones there?
In our particular field, the tests tend to be more in the CLIA certified lab developed test route. Even though it has been more than a decade for NIPT, and and it has become standard of care, recommended by medical guidelines, no NIPT is FDA-approved. The FDA approval does not give many advantages, and it can lock the tests down, limiting one’s ability to innovate and improve them as quickly.
In oncology, one of the biggest advantages of FDA approval is ADLT pricing. Once you get FDA approval, you qualify for advanced diagnostic pricing, which allows you to increase your average selling price for the test that you are providing while still allowing it to be accessible to patients.
While it is something that we have not done yet in either prenatal or oncology, we have plans to become FDA approved on both sides. On the prenatal side for our fetal antigen test, we have a phase 3 clinical trial with Johnson & Johnson for alloimmunized patients. These are blood incompatibility scenarios involving high-risk patients for whom a drug is being developed. It would be intended to be a companion diagnostic if the drug is approved and if we have FDA approval for it.
On the oncology side, an FDA approval can increase our reimbursement rate to our competitor levels through ADLT pricing, making them more profitable.
As the company grows, what leadership principles matter the most to you, in terms of keeping your product focused, keeping the science correct, and inspiring your team?
I wrote a founders letter as part of our S-1, and I think a lot of the principles there are what we want to continue to prescribe to. We talk about two things that are very important to us.
One is that pressure is a privilege. A lot of the time, we set few, but very ambitious, goals for our executive leadership teams. I think we have created this culture where, when someone is given a very difficult and ambitious goal, it means two things. It means it matters, and it means we trust that person to deliver it.
From that perspective, our executive team sees it as a privilege that they are being trusted to deliver an important, and difficult, goal for the company's success. We lay out for them how much of a difference it can make for our company, for our patients and for our providers. Seeing problems, or challenges as an opportunity is something that is very important to us. Throughout our past, until the past few months, we had not been profitable. Now that we are profitable, it remains very important to keep that mentality of setting ambitious goals and continuing to hit them.
The second is the concept of the ‘20-Mile March’. Jim Collins wrote about this in Great by Choice, which is required reading for our executive leadership team. The concept is that the companies that overperform other companies by 10x over long periods are not changing their strategy, or their hiring practices, or their growth targets based on the external markets.
I tell our executive leadership team that they do not have to watch the stock price. It does not matter exactly where it is because we are profitable. This happened with Amazon, after the dot com bubble. All the internal metrics were doing incredibly well, but the stock went down from around a hundred dollars to nine or ten dollars. Jeff Bezos said that they did not need to care about the stock price because, if they were doing the right things, they were building a generational compaafny, which turned out to be true.
IIt is very important to have that long-term ‘20-Mile March’ view. Given our differentiated and competitive products and our sales efficiency, many investors ask us why we do not hire hundreds of reps and scale even faster than our current growth rate, which is already around 100% YoY.
Our view is that if we did that, we’d break the quality of who we were hiring and the quality of service that we were providing. We are growing quickly, but if we try to grow even faster, we would break a lot of things. In healthcare, that is not acceptable. The service is critical. One might have great products, but if it is not coupled with great service, if the turnaround time is terrible, it is not going to help patients.
That is why we are growing fast while maintaining the quality of hiring and of service, hiring only people who are truly one percenters.
It is also important to note that when the market came down in the 2022 - 2023 period and a lot of diagnostics companies fell more than 90 percent because none of them were profitable, many of the same investors also questioned why we were still hiring. I had to build, and show, models explaining why it made sense to continue hiring, even during that period.
As a result of this view, we continued to hire and grow, and we were able to blow past our projections. As the market corrected, the environment improved and that continued growth allowed us to go public in a very strong position. If we had stopped hiring during that time, we would have lost years, as it is very difficult to re-accelerate growth.
Comments