By Sara Miller, NoCamels -
Welcoming a bundle of joy into your life should be a natural, happy event but millions of women around the world struggle to get pregnant and need a helping hand.
Since Louise Joy Brown became the first child born through in-vitro fertilization (IVF) in 1978, around eight million babies have been brought into the world using this method, in which an egg is fertilized by sperm outside of the body and then the embryo is implanted into the mother’s womb.
With Israel being the world leader in IVF procedures (in 2018 alone, the US Department of Health says it had almost twice the number of IVF cycles as the next country on the list, Japan), it is not surprising that a revolution in the efficacy of the treatment should also be Israeli.
In fact, while the global fertility rate has plummeted to an average of 2.3 per woman, Israel is the sole developed nation where it is above that number, with an average of 2.9 births per woman.
Tel Aviv-based AIVF, which uses AI to select the embryo with the best chance of being successfully implanted into a woman’s womb, was a startup co-founded by embryologist Daniella Gilboa following the completion of her PhD thesis on the topic.
Gilboa, who has worked as an embryologist for 15 years, is internationally respected in her field, with a role in global policymaking and membership in various committees of the American Society for Reproductive Medicine.
“I live and breathe embryos,” she tells NoCamels. “I am very connected to the ecosystem.”
In 2018, Gilboa joined forces with AIVF co-founder and reproductive medicine expert Prof. Daniel Seidman to create the company. Seidman, she says, is one of the busiest doctors in his field in Israel. Gilboa is CEO and Stanford-trained Seidman the chief medical officer.
“He brings the voice of the physician as well as the voice of the patient, and I bring the voice of the embryologist,” Gilboa says.
Working together, the concept of applying AI to boost the chances of women undergoing IVF began to crystallize.
“We started this to do better embryo evaluation, better embryo selection and really take out the human subjectivity,” she explains.
“Once you’re able to do this, you can really optimize success rates and the treatments, and this is the way to democratize IVF globally.”
Gilboa says embryologists like herself – who are usually experts in developmental biology – do have the expertise to evaluate the health of an embryo by monitoring different milestones and parameters throughout its growth.
But, she clarifies, relying solely on human judgment means that there is no standardization in the field when it comes to deciding which embryos fertilized in vitro are the most viable.
And, Gilboa warns, there is a “huge” range of the quality in clinics when it comes to individual IVF clinics and the services they offer, especially given that IVF is a worldwide phenomenon.
“You could have junior embryologists evaluating the embryos and really making the big $1 million decision of which embryo becomes a baby – this is in their hands,” she says.
Seidman and Gilboa began with the premise that an AI platform would be better than the human eye at the “crucial point“ of evaluating the viability of embryos in the lab and determining which of a woman’s fertilized eggs was most likely to result in a successful pregnancy.
Using “a huge amount”of biological data, Gilboa tells NoCamels, AIVF’s “EMA” platform was trained to understand developmental biology in order to detect milestones and parameters in the developing embryo, each of which correlates to the desired outcome, namely pregnancy.
EMA analyzes an embryo’s development in the incubator from “Day Zero” (fertilization) to about Day 5 or Day 6, just before it is potentially returned to the womb. After those five or six days, Gilboa explains, the AI produces a score for the embryo that translates to its probability of success.
This not only guides the medical professionals when deciding which embryo to implant, but also helps the patient understand the journey she is experiencing.
Gilboa says it has made a “massive” difference to the success rates of IVF treatments per clinic and, unexpectedly, has also led to a drastic reduction in failed pregnancies.
The company says EMA provides a 38 percent higher level of accuracy when compared to the evaluation of an embryologist alone. It also says that embryos with a high EMA score also demonstrate a 70 percent probability of a successful pregnancy.
Furthermore, Gilboa says, using AI helps move towards standardizing the process involved in selecting a viable embryo.
“It needs to be standardized across the IVF ecosystem, and this is what we’re trying to do,” she says. “It will get there.”
The company is now a commercial success, in use in Australia, Brazil, South Korea and a number of European countries, and soon to be launched in the US.
But as testament to its potential in the eyes of Gilboa and Seidman’s peers, in its infancy it received funding from IVF physicians.
“This was like endorsing the next generation of IVF,” Gilboa says.
Funding also came from angel investors (individuals using their own money, usually for early-stage startups) and venture capital firms, as well as in the form of grants from the Israel Innovation Authority and the predominantly European R&D program Eureka.
According to Gilboa, artificial intelligence has already been incorporated into medicine, with every physician having their own AI assistant. This makes the treatment process faster and more accurate, she explains, and allows physicians to improve their own performances with new tools to help make better decisions.
“And this is what we do in IVF,” she says.
