Medical technology keeps providing us with new information and ways to have a healthier pregnancy and baby.
This time, it’s a development that involves AI and pregnancy and could simplify and streamline in vitro fertilization processes, making the procedure less financially untenable.
This new technology would be able to examine embryos and assess them for specific traits related to viability.
In Vitro Fertilization Can Be An Arduous Process
In vitro fertilization, or IVF, is a process that can help some women get pregnant. For many patients, it takes multiple tries to achieve a viable pregnancy and bring a new human being into the world.
First, the patient is given treatments that increase ovulation for days or weeks; the patient then undergoes a procedure to harvest eggs from her ovaries. Then, the retrieved eggs are fertilized, and any resulting embryos are examined for viability. Embryos that seem likely to be viable are implanted into the uterus, and everyone crosses their fingers, hopes, and prays.
The whole process, which takes 4-6 weeks, according to the Cleveland Clinic, restarts if the embryos aren’t viable or fail to implant — which means that a couple may go through round after round of expensive treatments.
The New Technology Uses AI To Check One Aspect Of Viability
One key hint to an embryo’s likelihood of success is in the chromosomes, or more specifically, what’s called the “ploidy status.”
The embryo is examined for a number of complete chromosome pairs, and the new method would allow automated imaging tech to determine whether the embryo is likely to develop successfully. According to a new report in the journal Nature:
“We present BELA, a state-of-the-art ploidy prediction model that surpasses previous image- and video-based models without necessitating input from embryologists. BELA uses multitask learning to predict quality scores that are thereafter used to predict ploidy status.”
In short, BELA, which stands for Blastocyst Evaluation Learning Algorithm, examines a time-lapse video of the fertilized egg, looking for specific moments in development that allow the technology to assign a numeric score that helps predict the likelihood of successful implantation and live birth.
BELA Could Be The Next Step In Improving The IVF Process
Traditionally, the quality of an embryo has been assessed manually, with an embryologist using a microscope to examine it.
Then, there may be additional testing, particularly if there’s any specific reason (like maternal age) to have concerns about the embryo’s viability. Implementing new tech to improve this has been an ongoing process. Science Daily explains:
“The ‘gold standard’ test is a somewhat risky, biopsy-like procedure called preimplantation genetic testing for aneuploidy (PGT-A). In recent years, embryologists have been teaming up with computer/AI experts to find ways to automate some of this workflow and improve outcomes.”
BELA was trained on data from “nearly 2,000” embryos, and has been tested on existing datasets. The next step will be a controlled clinical trial.
If BELA is a success, it could make IVF a little more accessible and potentially save couples a lot of money — according to Fobres, a single round of IVF currently costs anywhere from $15,000 to $30,000.
Other Potential Applications Of AI In IVF
Since AI systems began developing more rapidly, there has been a lot of interest in their application in medicine.
In IVF, a 2021 paper published in the National Library of Medicine posited that in addition to assessing the quality of a fertilized embryo, AI might also be able to examine both egg and sperm cells to choose the healthiest ones for attempted fertilization. However, it also noted that this use might be limited because current practice is to fertilize all available eggs, and there isn’t currently strong data to use in predicting the most likely viable embryo and live birth based on examination of the egg.
Sperm cells, on the other hand, might hold more hope. The report notes that AI has already had some success predicting outcomes based on the examination of mouse sperm cell shape.
There Are Some Criticisms & Concerns
There are always concerns about introducing new technology into medicine. The hope is that new advancements will improve outcomes and that any negative impacts (foreseen and unforeseen) will be outweighed by the positive impacts.
When it comes to AI, the black box effect—the invisibility of the steps between input and output—is always a concern. According to a 2023 analysis in The Lancet, there are ways to diminish this effect, including making processes transparent and accessible.
The author also recommends that clinical trials be accompanied by long-term studies assessing the health outcomes for children produced via AI-assisted IVF.
What Does It Mean For Potential Parents?
It could be quite a while before AI is implemented into the IFV routine, although some lucky parents-to-be might get in ahead of the crowd by joining a clinical trial.
However, if and when it is successfully implemented, it could be a step toward making IFV more accessible and potentially more successful with fewer rounds of treatment and toward implementing AI in other medical processes.
There’s already AI tech available for diagnostics, transcription, and even remote treatment of patients. There are even programs that can identify malignant cancer cells, and AI used in screening can help reduce barriers to medical access, according to a 2020 article in the NLM.
Increasing applications of AI in medicine have potential drawbacks, but they may also make many diagnostic and treatment options far more accessible.
