01 Feb 2021 AI Improves Outcomes, Reduces Costs, in Assisted Reproductive Technology

Exciting new tools are using artificial intelligence (AI) to improve success rates for assisted reproduction technologies (ART) such as in vitro fertilization (IVF) and surrogacy, resulting in fewer procedures, lower costs and greater accessibility for more people.

One of the reasons I began to practice exclusively in assisted reproductive technology and fertility law was the excitement of the rapidly evolving field, as scientific advancements such as genomics, nanotechnology and cryopreservation continued to make parenthood a possibility for more and more people.

Technology, harnessed to serve the best interests of humanity, solves problems.

Technological advances have already improved outcomes for assisted reproduction, refining the in vitro fertilization (IVF) and implantation processes. Two decades ago, accepted practice was to implant multiple embryos to improve the odds that at least one would survive, resulting in a high incidence of twin and multiple births, with the attendant risks to babies and mothers or surrogates. As improved technology led to higher success rates, the industry best practice evolved to recommend implantation of a single embryo, significantly addressing the problem of multiple births with IVF.

The field of assisted reproduction faces other problems, including the high costs and scarcity of insurance coverage for fertility treatments such as IVF, which can put them financially out of reach for many people.

According to a 2019 National Institutes of Health report, “Artificial intelligence in reproductive medicine,” “The overall success of reproduction, either spontaneously of after ART, is highly dependent upon the quality of oocytes.” Currently, each egg retrieved via assisted reproduction has a 4.5 percent chance of resulting in pregnancy.

By improving the ability of IVF physicians to identify the eggs most likely to be fertilized and the embryos most likely to thrive, AI applications have the potential to minimize the number of IVF cycles. In the United States, medical costs for a single IVF cycle may range between $12,000 and $15,000, plus another $1,500 to $3,000 for medications. Reducing cost is one of the best ways to make assisted reproduction more affordable and thus more accessible, for everyone.

Artificial Intelligence Improves Sperm, Egg, Embryo Selection

Among the most exciting new advancements is the growing use of AI, specifically machine learning (ML) techniques, to improve outcomes in assisted reproduction. As reported recently in a Forbes column, a UK-Canadian team has launched VIOLET™, an AI computer algorithm that has the ability to predict egg fertilization with 77 percent accuracy and to predict blastocyst embryo development with 62 percent accuracy. The UK partner, CARE Fertility, found the program to be more accurate than its own embryologists, by 12 percent for egg selection and by 18 percent for blastocyst embryo selection.

One of the most promising applications of the new VIOLET program, CARE leading embryologist Dr. Alison Campbell told Forbes, “is for people who wish to preserve their fertility.

“For the first time, people who freeze their eggs can receive a report that not only includes images of each of their eggs, but also predicts the likelihood of them becoming a blastocyst-stage embryo (an important marker of their health) and helps indicate the ideal number of eggs which should be frozen to best preserve fertility.”

In 2020, Weill Cornell Medicine reported on a team of researchers there who created a “deep neural network—a class of AI algorithms modeled after the neurons in a biological brain,” dubbed “Stork,” then put it through a “deep learning” process, in which it was shown 12,000 photos of human embryos 110 hours after fertilization, which is the blastocyst stage. Each embryo photo was labeled “good” or “poor” based on embryologists’ assessments and actual pregnancy results. Following that training, Stork was able to assess additional new embryos correctly with 97 percent accuracy—far better than any individual embryologist on his or her own.

Another promising AI application to improve embryo selection was recently introduced by an Israeli startup, Embryonics, which hopes for U.S. Food and Drug Administration approval by the end of 2021, according to Mobihealthnews.com. The application, called Ubar, “uses geometric deep learning, the next generation of machine learning, in a clinical setting.” The application also is being used to create customized hormonal treatment for IVF patients and in selecting eggs for cryopreservation.

According to the NIH report, an AI application developed in 2017 is able to determine the motility of human sperm and identify abnormalities with 89.92 percent accuracy. Researchers were working on automated systems for grading the viability of human embryos as far back as 2012. AI also is being used to create prediction models that offer would-be parents an objective analysis of their prospects for attaining pregnancy as well as helping tailor treatment to improve outcomes.

AI Could Solve Problem of Abandoned Embryos

Yet another AI innovation uses facial recognition technology to help intended parents ensure their donor-conceived child will look like them.

Often intended parents who are using assisted reproduction hope for a baby who will look just like them. Exciting new platforms, such as Fenomatch, are using AI technology to identify donor egg or sperm most likely to result in a baby who bears similar facial features to the intended mom or dad.

If this technology is eventually developed to match the facial biometrics of blastocyst embryos, it might help to solve another big problem—the thousands of unused frozen embryos that result from assisted reproduction, many later “abandoned” in frozen storage, as we reported earlier.

Many clinics today offer intended parents, once they have created their families, the option to donate unused embryos, and the use of existing, donated embryos offers a lower-cost option for many infertile couples. However, many of the agencies offering embryo “adoption” are faith-based organizations that discriminate against LGBTQ and single intended parents, as we wrote recently The use of facial recognition AI in donor selection could go a long way toward encouraging the use of donor embryos… and over the long-run, could reduce the number of “abandoned” embryos.

AI Reduces Number of IVF Cycles, Improves Pregnancy Rates, Reduces Costs

Twelve years ago, when my spouse and I became fathers to twin boys via egg donation, IVF and surrogacy, we relied solely on the expertise and experience of our IVF providers, first to evaluate and select the healthiest eggs for fertilization from our egg donor, then following fertilization, to select the healthiest and most viable embryos for implantation. The process was not seamless. On our first IVF cycle, we ended up with three viable embryos, all of which were implanted. No pregnancy resulted. The second time around, we ended up with four viable embryos. This time, our surrogate got pregnant, and we became dads to twins.

We were lucky; we had the financial means to persevere and undertake a second round of IVF, with the attendant costs, in order to fulfill our dreams of parenthood. Many couples don’t have that advantage. But with the second IVF cycle, we, too, felt the financial pressure to implant multiple embryos, knowing that we might not be able to afford a third round.

If we were beginning our journey to parenthood today, lots of things would be different. With significantly improved outcomes resulting from technological advances, our IVF physician today most likely would advise us to implant only a single embryo. While there is no substitute for the knowledge and expertise of experienced IVF physicians and embryologists, some form of AI would probably be utilized to offer objective analysis and improve the odds of success.

Machine learning is still a new technology that is advancing by leaps and bounds. As more data on assisted reproduction selection and outcomes becomes available, the computer’s ability to assess and predict will become more and more accurate. Ultimately, technological advancements such as AI will remove the guesswork and human shortcomings from assisted reproductive technologies, making it safer, easier and cheaper for everyone who wants to become a parent to achieve that dream.

For information on how to ensure your family via ART and your parental rights are legally protected, contact the experienced fertility lawyers at IFLG today.