Artificial intelligence (AI) has been demonstrated to identify cancer, locate cavities, and provide medical information. Now, it may assist fertility specialists in choosing the best embryo for in-vitro fertilization (IVF).
According to the Centers for Disease Control and Prevention (CDC), one in five adult married women in the United States are still unable to conceive after a year of trying. As a result, many turn to IVF for assistance. 1% to 2% of all births in the nation are the result of fertility treatments.
The American Society of Reproductive Medicine estimates that the procedure costs an average of more than $12,000 each session and that it is not guaranteed. Additionally, most women need more than one try.
The odds are meant to be improved by Tel Aviv, Israel-based AIVF, a reproductive technology company.
Its AI-powered EMA software, which is used to evaluate embryos, is intended to analyse enormous volumes of data in order to make the embryo selection process simpler.
According to Daniella Gilboa, an embryologist and co-founder of AIVF, “IVF is one of the most significant medical developments in the last 50 years, but it’s not good enough.”
Only one in five pregnancies occur because success rates are only about 23% to 25% across all age groups.
The inability of IVF facilities to meet the rising demand, according to Gilboa, is one of the main problems.
The need for IVF is increasing, but there is a limited supply since more women are freezing their eggs to put off having children and concentrate on their professions.
Only 20% of the need is met in the United States, hence 80% of those women are giving up on having children.
The choosing of the embryo is one of the most important decision-making steps in the IVF process, according to Gilboa.
Human clinicians have traditionally been tasked with making that decision on their own.
Imagine being an embryologist and having to choose which embryo in a frantic lab setting has the best chance of developing into a baby, she added.
“You might have eight, ten, or twelve identical embryos, and you must occasionally decide which one to keep. In essence, you are observing the embryos beneath a microscope.
With the use of the EMA software for AIVF, an AI model assists the practitioner by assessing the embryos and using its sophisticated algorithm to select which ones are the best prospects for a healthy pregnancy.
“The AI has been trained to detect embryonic features that correlate to different outcomes — like genetic abnormalities, implantation, or gender — that can’t be seen with the human eye,” said Gilboa.
Each embryo is given a numerical rating. The ultimate choice might then be made by the clinician.
Gilboa stated that without AI, practitioners would have to rely on the embryo’s appearance to assess its quality — “but that’s based on subjective human analysis that doesn’t really quantify the actual chances of getting pregnant,” she added.
“The AI enables the doctor to reduce uncertainties and generate precise, understandable information.”
She added that the AI is also significantly faster, examining embryos in a small fraction of the time it takes a human, allowing clinics to treat more patients and handle a greater volume of demand.
Numerous time-lapse videos of developing embryos were used to train EMA’s AI program, which eventually learned which embryos would survive and which would not.
Gilboa added, “You need a lot of data. “To build a model that was good enough to start using took a long process, with many durations.”
Dr. Shahin Ghadir, a California-based Trimly fertility coach and double board-certified infertility specialist, said AI technology like AIVF might be “helpful and useful.” (Trimly is an internet resource for health care advice.)
According to Ghadir of Fox News Digital, “The field of reproductive medicine and fertility has experienced significant growth limitations due to the lack of manpower and expertise available to work in this industry.”
“This technology would be greatly appreciated because there are so few and difficult to train embryologists.”
Embryologists are still needed in laboratories, he continued, but merging their knowledge with artificial intelligence might be very beneficial.
Gilboa emphasized that although EMA aids in streamlining and speeding up the embryo selection process, it is not meant to take the position of medical professionals.
“At the end of the day, it’s the human who makes the decision, not the AI,” she remarked. It is merely a tool to aid in the clinician’s assessment of the embryos, consultation with the patient, and improvement of IVF service.
It’s “humans with AI,” not “humans versus AI.”
Ghadir underlined that before it is implemented, artificial intelligence must be “scrutinized under significant criteria” even though he supports assessing its potential.
“Some of the risks and limitations of artificial intelligence in an IVF setting and laboratory include errors in the grading of embryos that could potentially affect outcomes,” the expert stated.
The advice for employing AI in this situation is to compare it to expert evaluations side-by-side for a while, up until there is absolute certainty that the AI is just as accurate.
The idea of utilizing AI in embryo selection may potentially involve some ethical issues, which may have an impact on the individual choices made by patients and their families.
Currently being used in Europe, Southeast Asia, and South America, AIVF’s technology is anticipated to be made available in the United States “very soon,” according to Gilboa.
Feedback from the clinics who have bought and are utilizing the software has been favorable.
According to studies, IVF success rates at clinics have increased by 30%, Gilboa added.
She stated that because the technology might lead to successful pregnancies in fewer sessions, it might also potentially lower the cost to the patient.
According to her, pregnancy usually requires three to five standard IVF cycles, but only an average of 1.6 with AIVF.
Gilboa continued, “Through advanced AI, patients who desire a child can participate in the IVF procedure with complete laboratory transparency.”
