Victor Velculescu, DELFI Diagnostics — Founder Story

DFJ Growth
DFJ Growth News
Published in
13 min readFeb 16, 2024


You have to be innovative to create totally new ways of solving problems that can lead to a long-term solution.

Victor Velculescu, MD, PhD, has devoted his career to the fight against cancer.

One of the world’s leading experts in genomics and cancer research, Victor is the co-director of Cancer Genetics and Epigenetics at Johns Hopkins University School of Medicine, and his pioneering work has furthered the understanding of the molecular basis of cancer. But he may wind up leaving an even bigger mark in the field for what he’s done in getting those discoveries closer to patients outside of academia.

In his first try at entrepreneurship, Victor co-founded the cancer genomics company Personal Genome Diagnostics in 2010. The startup expanded access to genomic tests for cancer and was the first clinical laboratory to offer whole-exome sequencing. It was also the first to provide an FDA-approved kitted test for comprehensive genomic profiling for cancer patients. The company was acquired by LabCorp in 2022.

But Victor wasn’t done. He knew that the best way to reduce cancer deaths is through early detection, yet too many people still don’t get screened for cancer. Victor wanted to change that. So, he started his latest venture, DELFI Diagnostics, in 2019 to deploy machine learning to power a blood test (called a “liquid biopsy”) for early cancer detection. It’s a different approach toward liquid biopsies that Victor hopes will make a dramatic difference by providing affordable screening to everyone.

Following five years as the company’s chief executive officer, Victor has recently transitioned to a new role. He will remain as founder and a member of the board of directors, continuing to guide DELFI’s research and development portfolio for the future while turning the CEO reins over to Susan Tousi. Susan joined DELFI following a decade with Illumina, where she was most recently chief commercial officer. If the team succeeds, DELFI will be responsible for saving the lives of a countless number of people. Not bad for a kid who landed in the US when he was seven.

Q: You were born in Bucharest before the Berlin Wall came down. What stands out in your memory?

One of the things I remember vividly is that there was a massive earthquake in Romania in 1977. We were fine, but there was a lot of devastation. That might have been one of the earliest events in my life that got me thinking about a career where you work to help others, perhaps in a medically related field that can reduce suffering.

Q: What lines of work were your parents in?

My father was an architect, and my mother was a microbiologist.

Q: Those are good careers, but they decided to uproot the family and move to the US. That’s a big transition. Why did they decide to come over?

There were challenges in Romania at that time. You may remember there was a dictatorship, and it was very difficult to work and have the kind of opportunities we in the US view as commonplace today. They realized that both for them as well as likely for me, the future would have been limited in that environment.

Q: Where did the family settle down in the US?

We ended up in Westlake Village in Southern California. Parachuting from an Eastern Bloc country into Westlake Village was not something that I would’ve ever imagined. It led to a very idyllic and beautiful childhood for me.

Q: No culture shock? I would have imagined it was a little weird for a kid.

When you are born in a different part of the world, you have a different mentality that you bring with you. Of course, you’re learning the language and trying to fit in, but kids are very adaptable. I think my childhood certainly provided a unique perspective on how you view others with different backgrounds that has been very helpful as we are now becoming a global society.

Q: Did you always want to go into medicine? You mentioned the earthquake and how that affected you. But at what point in your life did you say to yourself that is something that you would like to do?

It was early on in high school that I first thought about medicine. We had family friends who were physicians, and I admired them and their impact in helping others. But it wasn’t until later, when I became an undergrad at Stanford, that I realized that I enjoyed not just medicine but the underlying science of it. It made me appreciate that if we understand disease better, we can make an impact that can have a huge effect on public health and society. That got me very excited because I understood that the value of what I could be doing could have an impact way beyond just the interesting aspects of science but on society as a whole.

Q: So, after receiving your MD-PhD from Johns Hopkins, how did cancer research become your passion?

It became clear in medical school that what I really wanted to do was utilize science in a way to understand human disease. Cancer came to the forefront, in part, because it’s one of the most complex diseases and affects so many people. It’s fascinating biologically because you can obviously make a gigantic impact. And there were great folks and mentors who I interacted with that helped lead me along this path.

Q: Being a professor of oncology at Johns Hopkins would keep most people busy enough. When did the entrepreneurial bug bite?

In grad school, some of the research my colleagues and I had been doing became of interest to diagnostic and pharma companies, and they began to license it. I thought, “Boy, this is so great. They’re going to take this and build great tests out of it.”

These large organizations would have great intentions. But they’re typically very bureaucratic and have many goals and many objectives before they get to your topic of interest. That got me thinking whether we can do things differently and organize this in a way that helps get ideas and, ultimately, products to patients faster. So, the first company that I founded was Personal Genome Diagnostics or PGDx. And the goal of PGDx was to translate our discoveries in cancer genomics to patients with cancer, helping get the right medicine to the right patient by better understanding their disease at a molecular level. It grew over time and was acquired about a year ago by LabCorp. Now, LabCorp has taken many of these tests and is providing them to patients worldwide through their various channels.

Q: Your background was in an entirely different field altogether. How do you retrofit that experience as a doctor into a commercial endeavor?

It’s not that different. Obviously, a company has a commercial goal, but you don’t reach that state until typically much later. The first part of starting a company involves identifying a problem and assembling a team, the right pieces of technology/IP, and the right sort of vision to solve that problem. That’s very tied to what we do day to day in our research endeavors in academia. I know what the problem is. I’m trying to get there through a certain vision, and I’ve got a team of folks who are working on that from a research perspective.

A small company can be very similar. You’re starting out with a new group and not immediately commercializing. You’re taking the science further to convert it into a product. Ultimately, it is different from what you do in a lab or a typical academic setting because there’s more engineering and product development. But you’re trying to get approaches that you develop as a prototype in an academic setting to be more honed and perfected, and then explaining the story to investors and others. The concept of teaching and explaining science and our vision was something we do all the time, and now I’m just doing it with investors.

Q: Interesting that those leadership skills transferred over. Were there also other skills that the job required you to develop along the way?

There are many that I’ve had to learn or adapt to this setting. One is that you have to be innovative to create totally new ways of solving problems that can lead to a long-term solution for patients. A second is that you’re trying to achieve a very practical objective, and you need to do this in a very clear, focused way. Sometimes, I think scientists try to make everything perfect. You obviously want the work to be based on very rigorous science, but things can’t take forever to make an impact. There’s the saying, “Don’t let the perfect become the enemy of the good.” I use that in my lab as well. There is a real risk that if the science takes too long or is unfocused, it will no longer be relevant. It’s something that I think is particularly appropriate for a commercial setting. You want to get to something that’s very robust and very good for patients but needs to get there quickly.

Q: Where did liquid biopsy fall short?

The best way to have an impact on reducing the mortality of cancer is by improving early cancer detection. For about a decade, we had been working in liquid biopsies as a tool to detect and understand cancer. We had several firsts, but none were good enough for what I thought was important for early detection on a massive public health scale. The first generation of liquid biopsies could help, but they were expensive and were not detecting enough patients with cancer because they weren’t sufficiently sensitive. The liquid biopsies in that first generation were typically used for already-diagnosed cancer patients to help detect recurrence or to select a treatment, which is like a million or so individuals. But they were not being used for the tens to hundreds of millions of people who needed screening nationwide or worldwide.

Q: Why weren’t they sufficiently sensitive?

The first generation of liquid biopsy tests typically were looking for specific changes or mutations. Mutations are some of the changes occurring in cancer, but they occur in a tiny part of the genome. Not a lot of the genome has mutations. And in many cases, when looking at a panel of a small number of cancer genes, a patient may not have any mutation. On top of that, we found that healthy individuals sometimes have mutations in their blood cells due to a process called clonal hematopoiesis. So that really highlighted the challenges of using mutations for cancer detection altogether. We needed to find a way to detect cancer in a higher-performing way yet in a way that was cost-effective and easier to do.

Q: When did that come about?

In 2018, some of my Johns Hopkins colleagues and I ended up having an “aha” moment. It focused on looking at a new way of detecting cancer based on fragmentation of DNA in the blood. We developed a new approach to compare the sizes and amount of DNA throughout the genome that you get in the blood. We saw that the DNA in patients with cancer was typically smaller or larger than those individuals who don’t have cancer. The reason for that is because DNA gets broken up into pieces when it makes it into the bloodstream.

The genome in normal cells is packaged in an organized way. When the pieces of DNA from healthy cells are cut up into small fragments, they’re done so in the same way every time while cancer patients have a disorganized genome. That leads to fragments of variable sizes from throughout the genome. And because there are so many fragments that we analyze, it allows us to have a much more sensitive test. In parallel, the laboratory test of getting the free DNA from blood and measuring its size through whole genome sequencing was done through elegant but straightforward laboratory steps.

Q: How did that aid your process?

Whole-genome sequencing allowed us to get a lot of information very inexpensively. What made it so powerful is that we layered on machine learning to take all this information of cell size and amounts throughout the genome and ask “How different is this profile from an individual without cancer?” It turned out that machine learning is very good at doing that. And we were able to distinguish between these two populations very nicely with this approach.

Q: So, you deploy machine learning to examine a broad spectrum of genomes from the bloodstream and then you compare DNA and look for changes between healthy cells and ones with tumors?

Exactly. In your blood, you’ve got lots of DNA fragments. They’re coming from genomes of normal cells and from genomes of cancer cells. In individuals who are healthy, pieces of DNA from throughout the genome have essentially the same profile. In patients with cancer, we’re finding that the profiles are different. In some regions of the genome, the DNA is bigger, and in some regions, it’s smaller. These changes, which are wide-ranging, are easily captured — in many cases by eye, and you can see the profile difference. But in other cases, you can see it very clearly with machine learning.

An analogy we use is of a pebble hitting the water. If you’re looking for a mutation to catch the cancer, you want to see that pebble, which is the mutation, right when it hits the water. If you’re there a little too early or late, you’re not going to see it. And in many cases, you’re going to miss it because it’s hard to see. In this setting, we’re not looking for the pebble; we’re looking at the ripples that occur in a pond after a pebble hits the water. And because you can see that in many different parts of the pond, you don’t have to be there exactly when it occurred. You can see it a little bit later.

Q: Talk about the platform DELFI developed and how it compares with the cost of previous methods.

While typical tests of this type can cost thousands of dollars, the (DELFI) test is likely to ultimately be hundreds of dollars because you’re using a very elegant approach in the laboratory, which has fewer steps. And then, on the sequencing side, we use what’s called low-coverage sequencing. The first-generation tests need sequencing of 10,000x or 30,000x of the same area. We’re only sequencing the genome at 1x to 2x, which basically means lower sequencing costs.

Q: By a factor of what can you bring down the cost for these tests?

It’s certainly going to be much, much lower than existing technologies. And that’s why we call this really a next-generation liquid biopsy approach because it provides more information for less cost. It’s a new way of doing liquid biopsies. One of the beauties of this approach is that it brings together new technology based on AI that I think will make a dramatic difference in the lives of individuals within the US and around the world — including those who would not normally be getting care because of the high cost. It’s one of these things where you’re making a big, almost science-fiction jump into the future.

Q: Where are you and your team in terms of deployment?

We’re developing the technology initially for populations at elevated risk of developing cancer. We have two national trials going on for individuals at increased risk of lung cancer due to their age and smoking history — one with 2,500 individuals and the other with 15,000 individuals. We’ve completed the initial part of the first trial to develop and validate our clinical lung cancer screening test called FirstLook Lung. The second larger trial is ongoing. The clinical value of our screening test is that it can be added to routine bloodwork and is highly sensitive, including for early-stage disease, with a negative predictive value of 99.7 percent. We’re currently introducing it to major US health systems through our Early Experience Program.

I’m excited to see health systems’ interest in FirstLook and how we’ll continue to expand access now that DELFI has reached an inflection point and Susan has joined to lead DELFI in this new phase. There’s a great need for lung cancer screening in many parts of the world, and this is a real human problem that needs to be addressed. This is the kind of disease that knows no boundaries. And because of my background, I think that I’m particularly passionate about seeing that this becomes as useful to as many individuals as possible worldwide.

Q: What do you like to do in the little spare time that you might have?

Good question. I love thinking about the future and how we might use these kinds of technologies to impact medicine. I have three kids and a family, and as they are getting older, I enjoy traveling and spending time with them. It has been really rewarding to see them grow up, as you can imagine.

Q: Before reaching the finish line, what’s your favorite book?

Lord of the Rings by JRR Tolkien. It highlights the importance of aiming to achieve almost impossible goals and how essential an incredible team is in achieving those goals.

Q: Favorite movie?

Star Wars. The idea of the future where we develop incredible technology with the goal of helping humanity is always something that has inspired me. Just thinking beyond where we are today to some world that we can imagine differently was always very moving.

Q: Is there a motto or saying that epitomizes or sums up the way you approach leadership and business?

There are a couple. One is “Don’t let the perfect become the enemy of the good.” This is something I’ve used both in the academic setting as well as at DELFI. The goal here is to create meaningful advances but to do them in a timely fashion.

We have spent some time thinking about values for DELFI. And in that process, we have developed three different ones that reflect underlying principles of leadership that I believe in. The first is “Lead with science, anchor in pragmatism.” You’re starting with science, but it’s got to be very rigorous science that can be practically useful. The second is “Build with and for all.” The concept here is that diversity is important — both in your team but also in who you think these tests will ultimately be useful for. The last is “Put we over I.” For any effort to be successful, you need to think from a team perspective. Obviously, we want everybody’s individual thoughts and contributions, but thinking as a team and thinking about a goal that’s larger than yourself often leads to everybody being able to do incredible things.

Q: Who is the one person who had the most impact on your professional career?

There are many individuals who have had a tremendous impact on my professional career, but too many to list. In terms of a historical figure who has inspired me and my work, I would say Louis Pasteur for his discoveries in medicine that have had an important impact on public health and in prevention of disease. His quote, ‘Chance favors the prepared mind,’ has been relevant for our work in discovering and recognizing key features of the cancer genome that may be useful for early cancer detection.