14 Mar 2023 AI Poised to Revolutionize the Future of IVF
In 1978, pioneered by gynecologist Patrick Steptoe and physiologist Robert Edwards, the world’s first IVF baby, Louise Joy Brown, was born in Manchester, England. The news of the first “test tube” baby made global headlines and at the time was likened to something out of a science fiction movie. Fast forward more than 40 years, and IVF has not only become a household name but also a mainstream medical procedure resulting in over 8 million babies having been born worldwide. As it has grown in popularity for intended parents struggling with infertility, an increasing number of employers in the U.S. are now offering fertility treatment coverage, and states are passing fertility health insurance coverage mandates as well. As a result, more and more intended parents are aware of their fertility options and benefits, leading to a drastic increase in the number of IVF cycles being performed.
Surprisingly, some systems in place for creating, storing, and choosing viable embryos have not changed much in 40 years, leading to an outdated approach unable to keep up with demand and other areas of fertility, which have improved drastically with modern technology. Now, newly developed Artificial Intelligence (AI) technology in the fertility world stands to change the way the industry operates these outdated systems and in doing so, may influence IVF success rates, the accuracy of cataloging frozen gametes, and lessen the overall cost of fertility treatments.
New Automated Robotics Helps Track Eggs and Embryos
Many cryobanks still operate by using handwritten labels and manual inventory checks and storage procedures and are only able to provide patients with limited access to specimen data. These outdated tracking systems create the potential for losses and mix-ups, and, although extremely rare, it does happen.
CNN reported in 2021 a mix-up in which two women received the wrong embryos during IVF and unknowingly carried and gave birth to children who weren’t theirs. Suspicions were confirmed after DNA testing proved the mix-up, but only after each woman cared for a baby that wasn’t theirs for several months.
Then earlier this month, FOX LA reported that a couple had launched a lawsuit with allegations of transferring the wrong embryo with a cancer-causing gene to an intended parent, resulting in a baby boy who now carries the deadly stomach cancer-causing gene. Both parents, who carry cancer-causing genes, had specifically sought out IVF instead of conceiving naturally so that neither gene would be passed on to their child. The mother tells FOX LA, “We wanted our children to not have any worry regarding these types of genetic mutations that we carry, so we tried to do everything in our power to give them a fighting chance at life with a healthy life.”
Thankfully, new automated technology is transforming the way frozen eggs and embryos are stored and tracked, possibly preventing any future losses or mix-ups. TMRW Life Sciences, a life sciences technology company, is using automated robotics to digitally identify and track frozen eggs and embryos while allowing patients the ability to monitor their eggs and embryos remotely 24 hours a day. The company gives instant access to embryo grading and egg maturity and even allows online payment for storage fees.
Artificial Intelligence May be Capable of Choosing the Healthiest Embryo
There is a multi-step approach before IVF can even begin. Eggs must be retrieved from the patient’s ovaries, and sperm must be collected. Eggs and sperm are combined to create embryos, which are then matured in labs for several days. Embryos are then graded based on appearance, size of their cells, and the rate of development, and if elected by the patients, embryos can be further tested for viability and chromosomal abnormalities (such testing is invasive, and it is generally referred to as “PGD” – short for pre-implantation genetic diagnostics).
According to IEEE Spectrum, however, there are several problems with this approach. David Silver, a machine learning engineer and one of the co-founders of the AI start-up Embryonics, tells IEEE Spectrum, “One is that the embryologists’ ability to collect data is limited. The amount of data about embryos, past patients, and successful live births available to any single doctor is very small, so it’s hard for them to generalize [about] what indicates that a fertilized egg is viable.”
As a result, many decisions for choosing a healthy embryo are solely based on each individual doctor’s experience with what worked and what didn’t work with past patients. Dr. Yael Gold-Zamir, CEO and one of the co-founders of Embryonics, tells IEEE Spectrum, “Many complicated decisions are made based on the doctor’s gut feeling, which is based on all the cases they have seen in their career.” Depending on their individual experience in the medical field, two embryologists may look at the same embryo and grade it differently.
With the development of AI algorithms, embryologists may now have some help in alleviating some of the guesswork when it comes to choosing healthy embryos. AI may also offer an overhaul of an outdated system that has not advanced in almost 40 years. According to National World News, researchers at Weill Cornell Medicine University have been developing an algorithm through AI that helps with healthy embryo selection with an accuracy of 70 percent. The algorithm determines whether an embryo has normal chromosomes or abnormal chromosomes, called aneuploidy, which aids in the selection of the embryo that has the best chance of implantation leading to a successful pregnancy. This method not only offers better accuracy but does not require a biopsy of the embryo, which not only adds cost to the IVF process but is also invasive to the embryo.
Dr. Iman Hajirasouliha, Computational Genomics, and Associate Professor of Physiology and Biophysics at Weill Cornell Medicine, and lead study author, tells National World News, “Our hope is that eventually, we can predict in a completely non-invasive way using artificial intelligence and computer vision techniques.”
Other algorithms classify images of embryos by comparing the patient’s medical information including age and health conditions and compare the patient’s data with past patients who had successful and unsuccessful implantations. The more data the algorithm collects from each patient, the more data it has to make science-backed decisions.
AI May Make IVF Cheaper and More Accessible to Future Intended Parents
A single round of IVF in the United State can cost anywhere from $12,000 to $15,000, and, with multiple rounds of IVF often needed to create a successful pregnancy, the compiling costs mean IVF is financially out of the question for many intended parents.
AI applications have the potential to minimize the number of IVF cycles with objective analysis and the ability to see differences in embryos that the human eye can’t see. As we reported in 2021, a fertility clinic in the UK tested out the AI algorithm VIOLET and found the algorithm’s predictions for healthy embryos to be 12 to 18 percent more accurate than their own embryologists. Higher accuracy for healthy embryo selection could reduce the number of IVF cycles needed for live birth, making it more affordable for everyone.
Using AI may also make Assisted Reproductive Technology (ART) more inclusive to the younger generation who wish to preserve their fertility. At present, ART is focused largely on patients struggling with infertility, but as AI continues to grow, Dr. Gerard Letterie, a reproductive endocrinologist and partner at Seattle Reproductive Medicine, predicts that this new patient segment will be interested in preserving their fertility by creating and freezing embryos for future use. He tells Forbes, “This will markedly expand the number of patients seeking care with assisted reproductive technologies.”
With more people seeking fertility care, the global IVF market is expected to reach $36 billion by 2026. Dr. Letterie goes on to tell Forbes that he predicts that there “simply won’t be enough embryologists to address the rising demand” for ART. With a possible embryologist shortage, AI may be able to help with the growing demand by helping embryologists determine viable embryos faster and with more accuracy, which can maximize successful outcomes and reduce the time it takes for intended parents to conceive.
As AI paves a new path in fertility treatment, opportunity rises for a more diverse group of people who may not normally have access to care. AI will ultimately take the guesswork out of assisted reproductive technologies by providing science-backed data, removing human error, and in the process making services cheaper and safer, giving everyone, regardless of geographic, racial, or socio-economic background, an equal stake in the fertility world.