A new artificial intelligence (AI) algorithm developed by an Israeli startup could help improve the chances of pregnancy for millions of couples who undergo in vitro fertilization (IVF) treatments. The algorithm, created by Embryonics, uses machine learning to automatically analyze embryos and select the one most likely to result in a successful implantation.
How the Algorithm Works
IVF is a process that involves fertilizing an egg with a sperm outside the body and then transferring the resulting embryo into the uterus. However, not all embryos have the same potential to implant and develop into a healthy baby. Currently, embryologists rely on subjective grading systems and time-consuming techniques to choose the best embryo for transfer.
Embryonics’ algorithm, called Ubar, aims to overcome these limitations by using data from tens of thousands of IVF cycles, including time-lapse videos of embryos. The algorithm learns from the outcomes of previous cycles and predicts the implantation potential of each embryo based on its size, shape, and other morphological features. The algorithm can also recommend which embryos not to use, which can save costs and prevent miscarriages.
The Clinical Results
Embryonics has conducted a clinical study at the Nadiya Clinic of Reproductive Medicine in Kyiv, Ukraine, to test the performance of its algorithm. The study involved 608 blastocysts, which are embryos that have developed for five to six days after fertilization. The algorithm outperformed a panel of human experts in predicting which blastocysts would result in pregnancy by about 20%. It also outperformed the human experts by nearly 30% in recommending which blastocysts not to use.
The study reported its first six successful pregnancies using the algorithm, while five other participants are awaiting test results. The company expects to get clearance in Europe in the next few weeks to begin using its technology in hospitals. In the United States, the company plans to conduct a clinical trial at several sites with the expectation of obtaining approval from the Food and Drug Administration by the end of 2021.
The Implications for IVF Patients
According to Embryonics, its technology could significantly increase the success rate of IVF treatments, which currently stands at about 30% in the United States and even lower globally. The company estimates that its algorithm could improve the overall success rate of an IVF process by an average of 10%, which would result in more “AI babies” and less emotional and financial stress for couples.
The company’s CEO, Dr. Yael Zamir, said that they are not trying to mimic the human expert, but rather introduce a new level of accuracy and objectivity. “We are using AI to help doctors make better decisions and provide better care for their patients,” she said.
Embryonics is not the only Israeli company that is using AI to improve IVF outcomes. Another startup, Fairtility, has developed an AI algorithm that can help improve the chance of pregnancy in IVF by automatically analyzing embryos as well. Fairtility’s algorithm was used in a study conducted by Ben-Gurion University’s School of Public Health in collaboration with several medical centers in Israel. The study found that blastocyst size was significantly associated with implantation potential and that larger blastocysts had higher odds of implanting in the womb.
The use of AI algorithms for embryo selection could be a game-changer for the field of fertility treatments, which has been struggling to increase its success rates for decades. By using data-driven and objective methods, these algorithms could help millions of couples achieve their dream of having a child.
