Advances in artificial intelligence (AI) are poised to revolutionize several areas of medicine, starting with diagnostics. Several studies have thus reported that AI could sometimes do better than the human eye. when it comes to diagnosing canceror to analyze a dubious mole.
The assisted reproduction could also benefit from this technological progress, according to a new scientific study. Published online this December 19 in the magazine The Lancet Digital Health (Source 1), it reveals that artificial intelligence could determine with approximately 70% accuracy whether a fertilized embryo in vitro is genetically viable. Hear if it has the right number of chromosomes, or if it has too many or too few.
Aneuploidy, i.e. having an abnormal number of chromosomes, is one of the main reasons for in vitro fertilization failures. Since the embryo is not genetically viable, it fails to implant in the uterine lining, or results in a miscarriage. To detect aneuploidy, one of the methods currently used is embryo biopsy: one or two cells from the embryo are removed for genetic analysis. It is this same technique that is used in case of pre-implantation diagnosiswhich is not authorized only in the event of a hereditary genetic disease in France. Most often, the medical teams of assisted reproduction centers choose the embryo(s) to be implanted according to their appearance under the microscope, their rate of cell division, their morphology. Because in addition to the ethical aspect of a possible genetic analysis, it also has a significant cost and remains an invasive act.
A way to reduce costs but also the number of embryos to be implanted
Here, the new algorithm, named STORK-A would help predict aneuploidy without the drawbacks of biopsy. Designed in 2019 by the same research team, this algorithm works by analyzing the images obtained under the microscope, and cross-checking these with other information, such as the patient’s age and how the ART center scores the embryos. Like any artificial intelligence, it “learns” as it goes, by correlating certain data with the risk of aneuploidy. In all, during its formation phase, the AI screened 10,378 blastocysts (the name given to embryos 5 days after fertilization) whose chromosomal status was known to scientists.
The team of researchers then found that the AI was able to predict aneuploidy with an accuracy of 69.3%. Regarding so-called “complex” aneuploidy, involving more than one supernumerary chromosome, the AI managed to identify them with an accuracy of 77.6%.
“It’s another great example of how AI can potentially transform medicine. Algorithm transforms tens of thousands of embryo images into AI models that can ultimately be used to help improve IVF efficiency and further democratize access by reducing costs”, enthused Dr. Olivier Elemento, co-author of the study and biomedicine researcher at Weill Cornell Medicine (New York, USA). “We believe that ultimately, by using this technology, we can reduce the number of embryos to be biopsied, reduce costs, and provide a very good consultation tool with the patient when they have to make the decision whether or not to a pre-implantation diagnosis”, added Dr. Zaninovic, also a co-author, in a press release (Source 2). This new approach could also make it possible to reduce the number of embryos implanted, and thus reduce the risk of multiple pregnancies that results.
It remains for researchers to continue the tests so that this artificial intelligence reaches an accuracy of more than 90%and that this approach be compared with conventional pre-implantation diagnostics in order, eventually, to replace them.
We would love to give thanks to the author of this write-up for this remarkable content
IVF: artificial intelligence could help choose the “best” embryo
Take a look at our social media accounts as well as other related pageshttps://yaroos.com/related-pages/