From Research to Entrepreneurship: How One Start-Up Is Working To Optimize IVF Success Rates

Many couples struggling with fertility problems pursue in vitro fertilization (IVF), a procedure that involves the fertilization of an egg outside of the body. While the treatment has resulted in the birth of over eight million babies worldwide, the success rate remains low. The life science company ImVitro hopes to change that, by developing tools to optimize the performance of IVF.

Technology Networks recently had the pleasure of speaking with ImVitro’s CEO and founder, Dr Alexandra Boussommier-Calleja, to learn more about IVF and how artificial intelligence (AI) could help to address some of the challenges associated with the technique. In this interview, Alexandra also discusses her journey to create ImVitro and provides some tips for researchers looking to build their own start-up company.

Anna MacDonald (AM): How does IVF work? What are some of the key considerations and main limitations of the technique?

Alexandra Boussommier-Calleja (AB): IVF is a medical treatment that aims to give help to people who are struggling to have children. Infertility today is thought to affect one in seven couples worldwide: it is very unlikely that you do not know someone yourself that has needed IVF or who is born through IVF.

IVF involves manipulating gametes to generate embryos which are maintained alive in the lab for a few days and assessed prior to a potential transfer back into the patient. IVF has made huge strides since its onset and has helped many patients; its success rates can still however stagnate at around 20-30% in Europe, which implies that many patients have to repeat this expensive, invasive and emotionally difficult procedure. There is a myriad of challenges with IVF, one of the main ones being how multi-factorial it is; in some ways more than for other medical treatment given that we are not just talking about one patient, given that we use gametes from different people.

Much like other medical treatment, IVF is challenging because it requires a series of complex medical decisions, each of which can impact the success rates and each of which should be personalized to the unique patient’s needs. Today, while already helping many patients, clinicians lack the tools to make some of these key decisions in a data-driven, objective and robust way. One of these decisions is choosing the embryo that has the highest chance of leading to a pregnancy and a birth and that should be transferred back into the patient. Embryologists today spend significant time evaluating the embryo development through microscopes and use their undeniable expertise to look for biological hints. Yet, so many things happen visually, that it is now clear to the scientific community that we are probably missing out on hidden clues which could save some patients from going through unsuccessful treatments. This is where ImVitro comes in.

AM: ImVitro’s vision is to “create an easier path to parenthood”. Can you explain how combining AI approaches and cell culture can help to achieve this?

AB: At ImVitro, creating an easier path to parenthood consists in minimizing the number of unsuccessful treatments patients have to go through to become parents. To do this, we want to develop tools that can learn from a myriad of parameters that characterize the journey of IVF patients. We also want these tools to detect hidden patterns that might have gone unnoticed so far by experts. Machine and deep learning algorithms are uniquely positioned to help us get there, as they can ingest huge swathes of data, and, if trained properly, learn as objectively as possible what is predictive of a pregnancy and birth, without “limiting” itself to the current clinical expertise. On the other hand, we also want to make sure we take into account the variability associated with cell culture and imaging, which in my experience is never negligible. This entails not just focusing on images, but also on contextualizing clinical factors that could help experts obtain more reproducible results across patients and clinics such that success chances are maximized for all patients.

AM: Can you tell us about the history of ImVitro and some of the key steps in its development?

AB: As a researcher I spent a lot of time in the lab, working with in vivo, ex vivo and in vitro pre-clinical models, with one common theme: microscopy. We’d spend ages preparing our assays, and after imaging them I felt all too often like I was missing out on key information that I could not or did not know how to extract objectively. I also often felt that there was a lot of variability in my results and imaging, probably stemming from uncontrolled cell culture conditions. With these core themes in mind, some of which I explored even further through some articles I wrote, I set out to develop tools that can automate this process and improve quality control for cell culture and cell imaging.

I obtained a Marie Curie fellowship tailored to help researchers integrate European start-ups and adapted my project to starting my own company. I joined Entrepreneur First at a stage where I had a beginning of an idea, but very little traction: here we focused solely on customer development, validating traction and fundraising to accelerate the process.

Nine months later, after obtaining a grant, I secured a seed round with international investors. This happened right as the COVID-19 pandemic started, but thankfully I had already secured strong partnerships and some data we could start working on with our first data scientist that I hired following the fundraise. More than a year later, I have built a team of seven soon to be eight people and we are about to release our first CE Marked product to our partners to perform pilot studies to keep improving our software with their help. Needless to say, it has been and still is an amazing entrepreneurial but above all human experience!

AM: ImVitro is backed by Entrepreneur First, an international talent investor. How important is support such as this to start-ups?

AB: There are endless ways to start a company, as a function of your project, expertise and existing support system. I can only comment on the two ways I tried to start my company: as an “intra”-preneur and as part of Entrepreneur First. When I was trying to start my company using the Marie Curie Fellowship at Elvesys, I did have some support and interesting advice from the team and could get direct and valuable technical help from some team members. The founders also insisted very much on the fact that I should spend significant time talking to potential customers. That being said, I was still having to force myself outside of my comfort zone, which is harder than it sounds. When you are a researcher, calling up people to do market research and to look for hints that will probably prove that your idea is wrong is not necessarily natural. Joining Entrepreneur First proved to be a stepping stone that decisively forced me outside of my comfort zone: I was not in a labor an R&D environment anymore and was incentivized to literally spend time on only one thing: talk to customers, and find traction. This to me was incredibly helpful, especially when surrounded by 50 other people whose sole concern is similar and who go through the same ups and downs. Knowing that you are not the only crazy person to try to kick-start your idea from the ground was often reassuring. Besides, the whole Entrepreneur First support network was useful in hearing some hard truths; you don’t have to listen or believe them all the time, but at least they are here to challenge you in a way that few other people would!

AM: Do you have any tips for other researchers looking to build their own start-up?

AB: There is no one-way to start a company, and I am still learning as we “speak” as to how to build a long-lasting successful start-up. My only tips at this stage would be:

·         To first spend most of your time looking for a co-founder if you’ve decided you can’t or don’t want to do this alone. Finding co-founders is far from being easy and can take months if not years. Make sure you pair up for the good reasons and work together on brainstorming and customer development before officially moving forward to directly test how you work together, and make sure you have hard conversations – they will come up, the longer you wait, the worse they can get.

·  To then spend all of your time doing customer development so as to identify a problem. I’m often told that the hardest part is to find a real problem: you can then “more easily” come up with a solution. As a researcher I find that this is very hard as we often (of course not always) push a technology that we hope will fill a gap. Here, we have to turn the problem on its head, and not be tempted to push our technology nor our expertise too, but rather first find a need that pulls it.

· And then, to aggressively ask for help: whether it is looking for accelerators such as Entrepreneur First, or to discuss your doubts with other founders or would-be founders etc. I find that the hardest part of this job is to be a continuous source of energy and trust for every stakeholder around you, and as a founder you will have tons: your investors, your employees, your customers… This is not a sprint but a marathon. Take care of yourself!

· And perhaps most importantly, find a problem that you can devote yourself entirely to. You will be thinking about your company day and night, so you absolutely need to be passionate about it! This is another reason why I started this company: working on IVF is not just scientifically mind-boggling, it also has many ramifications that are incredibly interesting (legally, ethically etc.) to explore on a day-to-day basis.

Dr Alexandra Boussommier-Calleja was speaking to Anna MacDonald, Science Writer for Technology Networks.