AI to tackle infertility: how machine learning is supporting the path to parenthood

In vitro fertilisation (IVF) has become a source of hope for many prospective parents encountering fertility issues. However, the process is not perfect, often requiring several cycles, which can be invasive, expensive and emotionally demanding, with no guarantee of success.

Alexandra Bousommier, CEO of ImVitro, is increasing the odds in patients’ favour.

“We’re using AI to tackle infertility.

ImVitro is developing machine learning algorithms that help IVF doctors and embryologists to make better predictions about which embryos have the highest chance of leading to a pregnancy and a birth,” she explains.

The company, founded via Entrepreneur First, aims to improve IVF success rates, shorten the time-to-pregnancy and time-to-birth for prospective parents, while saving embryologists time, via a unique SaaS platform.

While we’ve recently been speaking to the female leaders of companies using technology to improve the lives of women in our Focus on FemTech series, Alex’s company is different. Despite many perceiving her company as FemTech or female focused, she’s determined to challenge the perception that neither fertility nor the work that ImVitro does is a gendered issue.

Here, Alex shares her journey from researcher to CEO, including how she discovered the problem ImVitro solves, why right now is the right time for the company to exist, and why she wants people to think differently about gender and IVF.

Alexandra is the definition of a technical CEO. A biomedical engineer, Alex has a PhD from Imperial College London and a postdoctorate from Massachusetts Institute of Technology. It was there that she first encountered the problems that would become the foundation for ImVitro.

“I’m a researcher at heart. While I didn’t directly work directly in embryology, I did a lot of work with cell culture – the techniques that help you maintain cells alive in the lab, outside of the body. 

I spent a lot of time in the lab using microscopes and trying to understand visually what happens with cells and tissues. However, I couldn’t help but notice that there was a huge lack of automation in my field. That was tolerable in the realm of academia and fundamental research, but I doubted it could work for very long once you got closer to patients. 

Real patients need real solutions fast – so once you move past research, there’s a strong need to introduce automation and almost more importantly, a robust quality control.

With these thoughts heavy in my mind, I received a Marie Curie Fellowship and came back to Europe to use it as a stepping stone to kick-start my own company. That’s when I discovered the world of startups, and the desire to identify an addressable and pressing clinical need that I could entirely devote myself to.”

After exploring several avenues on her own, and maturing her ideas by writing articles and interviewing cell culture experts for a scientific editor, she joined Entrepreneur First’s Paris programme to accelerate her entrepreneurial adventure 

“EF were really quick to encourage me to look for traction less geared towards fundamental research, and more geared towards current clinical needs.

After testing several hypotheses related to varying cell culture needs, one day, the need became obvious. I realised that at the heart of IVF, you have cell culture, and with it, many of the same issues I had encountered during my research, including a lack of tools to analyze images of embryos.”

Utilising her expertise in cell culture, and the difficulties with automation she knew well, Alex contacted many embryologists to understand their workflow .

“Let’s imagine a couple who are embarking on their IVF journey. Currently, you generate around 5 embryos for them. 

An embryologist then has to maintain these embryos alive in the lab for around 4 to 6 days, at which point they choose which one to reimplant back into the patient. 

At the moment, these are assessed semi-qualitatively. Embryologists look for biological hints to gain a sense of which embryo has the highest chance of leading to a pregnancy.”

Here, Alex discovered a pain point. 

“Embryologists were generating huge amounts of images and videos of developing embryos, but didn’t have the tools to make the most of them. These videos had something to tell us, but we weren’t giving them a voice and were probably missing out on key information that could dramatically change the journey of many IVF patients. That’s where the dots connected.”

Alex had her problem, and set about building ImVitro to solve it.

“We’re taking a SaaS-driven approach towards building predictive models around biological data.

We help embryologists make the most of their videos and their clinical data by making predictions using our machine and deep learning algorithms. These algorithms were trained on retrospective data for which we know there was, or not, a pregnancy and birth so as to remain as objective as possible in our predictions”

Alex believes that now is the perfect time for her to be building her company.

“We are at the intersection of several factors that mean it’s the right time to do this; not five years ago, and not in five years.

The first is that the push to choose only one embryo is actually quite recent. In the past, and still very often today, you would see labs that would choose to reimplant two or three embryos in the hope that at least one leads to a pregnancy. 

We now realise that multiple pregnancies can be dangerous, both for the mother and the children, so it’s better to avoid it.

Secondly, the microscopes that record the embryo as it develops have only recently become more common in IVF centers. Before, to have a sense of how the embryo develops, you’d have to take it out of the oven maintaining it in physiologically relevant conditions, look at it under the microscope and put it back – all extremely quickly.

This was problematic for 2 reasons.

Firstly, embryos would suffer from a change of environmental parameters like temperature, and that’s really not good. Secondly, you’d be judging the development of the embryo based on a very quick look at it. And you don’t need to be a biologist to imagine that you’re missing out on a lot of information when you do this.

It’s really only in the last few years that more and more clinics have gotten this new microscope that can monitor the embryo continuously. Now, we are in a good position to make the most of these videos. 

Another key factor consists in the advent of vitrification in IVF (the practice of freezing an egg or embryo with extremely rapid cooling). Recent advances have made it much more successful, and it is therefore routine to freeze eggs or embryos for a potential future transfer. This means that we do not have to put all our eggs in one basket (pun intended) anymore with the first transfer, and can feel free to choose between different embryos and different timings.

This brings me to the fourth factor, which is the advent of deep learning and artificial intelligence.

Up until recently, it was taken for granted that it would be really hard to do what we’re doing. But now, we have the computer power to process huge swathes of information and the medical field is ripe for this technology: I am convinced we’re in a strong place to support this specific application.”

With the technology behind ImVitro’s first product almost completed, the company are now preparing to release to their partners the first version of their product next month

“I’ve raised a seed round with several VC funds in Europe and in the US, and have built a great team. We’re curating the data we have, and we’re using it to train the algorithms. We’re now finalising the research and development of the first algorithms, the development of the SaaS platform interface and the documentation required to obtain our CE Mark [certifying that the product complies with EU health, safety and environmental requirements]. 

This means we are now in a unique position to officially launch pilot studies with our partners, to iterate with them to develop the best possible product and science.” 

However, the journey from an idea to the Doctor’s office has not been simple.

“As what we’re creating is a medical device, iterations are not easy – you’re talking about real patients who will be genuinely impacted by our technology, so you cannot get a new version out every month; every minor or major change requires more or less heavy documentation or even audits from regulated bodies

Doctors are interested, but understandably skeptical, because AI can be seen as a ‘black box.’  There’s a challenge in showing them that it’s not something that they have to blindly trust, and in being as transparent as possible into how you have trained your algorithm so they can decide for themselves if it fits into the way they would reason and make a decision themselves. They’re eager to be shown that it has an impact.

This then presents another difficulty for us: while there’s a lot of value in being able to demonstrate the impact of your work, we’re talking about pregnancy and birth, which takes 9 months. You need quite a bit of time to show an increase in KPIs. This is probably going to be one of the next big hurdles for us.”

The ethical implications associated not only with IVF, but the technology behind their product, are another area that Alex and her team have to be conscious of.

“IVF is an amazing subject to work on, because other than being scientifically mind boggling, it has this huge societal and ethical dimension”, reflects Alexandra.

“It’s understandably a sensitive subject in some ways. On one level, there’s simply the fact that some people don’t agree with it, perhaps for religious or other personal reasons. 

On another level, there’s the question of, by helping doctors to make these decisions, are we conscious of the bias we might be introducing, and that an AI algorithm can have? 

We need to make sure we’re not introducing more bias into the evaluation of embryos, and as much as possible, that we are not automating any already existing bias. To address this, we’ve been intent on getting a diverse source of data from different clinics, with labels that required as little human intervention as possible. It’s incredibly important to think about how you source, and how you label, your data.”

Have these issues affected the way ImVitro raise capital? “Investors are becoming more and more aware of these issues, and I believe they will become gradually more challenging. Because they would see the impact that the product could have, this was not however a blocking issue as opposed to something that needed to be addressed from the onset of the company.”

We first approached Alex to take part in this interview for our ‘Focus in FemTech’ series. However, while some might see fertility as an issue centred on women, Alex considers it vital we collectively broaden our view, recognise that infertility is an issue that affects all genders, and challenge this perspective.

“I very often hear, ‘of course, the CEO of your company should be a woman. It’s an IVF company’,” Alex says. “It’s so not true.”

The misconception comes up so often that ImVitro’s onboarding booklet has a section on gender and IVF for employees to increase awareness, and be ready to challenge when it comes up.

“The idea that IVF relates more to women than men is wrong for several reasons. 

The first and most obvious one, is that infertility affects both men and women. They typically say that in a heterosexual couple, in one-third of cases, the infertility comes from the woman, in one third of cases the man, and in one third, it’s either from both or unexplained [1]. So factually, infertility is never going to be just a woman’s problem. 

The second one is that by implying IVF and fertility is a women’s issue, it suggests it should be more important for women to have kids than for men. This can be really detrimental to women – women can even be seen as having failed in their role if they struggle to or can’t have children. Equally, when we assume that fertility is a woman’s issue, we ignore the emotional pain experienced by men who want, but are struggling, to become fathers. 

Finally, IVF is not just used by heterosexual couples, but by same sex couples too. Therefore in suggesting that IVF is a ‘women’s issue’, you neglect the experiences of a huge set of people undergoing, or who have been through, IVF.

Ultimately, it would be doing us a lot of good if we just admitted that IVF and fertility is an issue affecting all genders. That being said IVF is not medically equal to all, given that retrieving sperm in sufficient quantities is generally dramatically less invasive and painful than retrieving enough ovules, let alone carrying the pregnancy.”

At the core of the company for Alex is remembering the real, human impact of imVitro’s technology, and what it can mean to the patients they serve.

“It’s important for us a company to be empathetic to all of the social and emotional issues, and inequalities, surrounding IVF. For example, it is also very important for us to recognize that not everyone wants to have children, and that we are here to help those who want to become parents and not to encourage everyone to be parents; all too often, childlessness tends to be equated to infertility, but it can very well be voluntary. 

When I work with my team, I often try to bring the focus back to who we’re helping and the big picture of this life-changing company – and to be honest, I believe this is a big reason for them to have joined ImVitro to start with. I spend extra time, especially at the beginning when onboarding, explaining the journey of the patient – because even though our customers are doctors, we’re working for patients, to address a huge and pressing societal issue. It is hard nowadays to not know a single person that has gone through IVF. They, and the difficult IVF journey they described to me at the very beginning of this adventure are the reason we do what we do.”

For more information on ImVitro, visit im-vitro.com

[1] Source: NICHD