This startup says its AI can better spot a healthy embryo — and improve IVF success – TechCrunch

With each year, AI is beginning to bring more standardized levels of diagnostic accuracy into medicine. this is true Skin cancer detection, For example, and lung cancer.

Now, a startup in Israel is called Embryology It is said that its AI can improve the odds of successfully transplanting a healthy embryo during in vitro fertilization. What the company is developing is, in essence, an algorithm to estimate the likelihood of embryo implantation, one that they have trained through IVF time-lapped imaging of developing embryos.

It’s just starting, to be clear. So far, a pilot has included 11 women between the ages of 20 and 40, with six of those individuals enjoying successful pregnancies, and the other five awaiting results, Embryonics says.

Nevertheless, the fetus is interesting for its ability to shake up a large market that has been stuck for decades and only continues to grow due to external trends, such as millennial women who are having children. Due to economic concerns.

Consider that the global in-vitro fertilization market is expected to grow from around $ 18.3 billion to almost double in the next five years Some guesses. Still tens of thousands Every year women undergoing IVF face costs from anywhere for a long time. $ 10,000 to $ 15,000 Per cycle (at least in the US), together Long shot odds Which worsens with age.

In fact, it is likely to reduce the number of IVF rounds and their attendant expenses for those who drive Embryology, Founded three years ago by CEO Yale Gold-Zamir, an MD who studied general surgery at Hebrew University, yet became a researcher in an IVF laboratory with an interest in the science behind fertility Were the reason.

As it happens, it will be offered to two individuals with complementary interests and expertise. One of them was David Silver, who studied bioinformatics at the prestigious Technion-Israel Institute of Technology and who spent three years as a machine learning engineer at Apple and before that at Intel, before joining Embryonics last year Three years were spent as an algorithm engineer. .

The second person to be introduced to Gold-conscience was Alex Bronstein, a serial founder who spent years as a lead engineer with Intel and who today heads the Center for Intelligent Systems in Technology as well as in-depth learning Has been associated with many efforts. Including Eric, embryology and so on Sibylla A.I., A nascent dress focused on algorithmic trading in capital markets.

It is a small organization, but three, with 13 other full-time employees joining them, appear to be making progress.

A $ 4 million share in seed funding led by the Schuterman Family Investment Office (led by Sender Cohen, former president of Soros Capital) and the Israeli Innovation Authority, Irionics says it is about to receive. regulatory approval In Europe that will enable it to sell its software – which it says the team can recognize patterns and interpret the image in small cell groups with greater accuracy than humans – fertility clinics across the continent.

Gold-Zamir says that the company is already tracking the next steps, using databases with millions of (anonymous) patient records from different centers around the world.

Most notably, beyond analyzing which of the many embryos is most likely to flourish, embryogenesis seeks to work with reproductive clinics to improve hormonal stimulation, so that their patients produce mature eggs as much as possible.

As Bronstein puts it, every woman who goes through IVF or fertility protection undergoes a hormonal stimulation process – which involves receiving injections with hormones from 8 to 14 days – to produce multiple eggs to her ovaries to inspire. But right now, there are just three common protocols and “a lot of trial and error in trying to establish the right test,” he says.

Although deep learning, fetalism thinks that it cannot begin to understand what hormones each person should take but at different times.

Gold-Zamir says that in addition to embryo selection, embryology has developed a non-invasive genetic test based on analysis of visual information, which in some cases can detect major chromosomal aberrations such as Down syndrome.

And there is more in the works when everyone is employed. Gold-Zamir says, “Ambryonics aims to provide a holistic solution covering all aspects of the process, raising four children of their own as they run the company.”

It is too soon to say whether the newborn dress will succeed naturally. But it is certainly at the forefront of a technology that is rapidly changing after more than 40 years, with many IVF clinics worldwide reporting their cell multiplication by looking at day-old embryos on a petri dish under a microscope. To assess is the health of the fetus. And size.

For example, in the spring of 2019, investigators at Weil Cornell Medicine in New York City Publish your own findings AI to assess an embryo morphology more accurately than the human eye after using 12,000 photographs of human embryos taken exactly 110 hours after fertilization to train an algorithm to discriminate between poor and good fetal quality. Can do.

Investigators stated that each embryo was first assigned a grade by the embryologist that considered various aspects of the fetus’s appearance. The investigators then performed a statistical analysis to correlate the grade of the fetus with the probability of a successful pregnancy outcome. If more than 5 percent and poor quality is less than 35%, the fetus was considered good quality.

After training and validation, the algorithm was able to classify the quality of a new set of images 97% accuracy.

Photo courtesy: Tammy Bar-Shay