2023-11-19 08:03:59
The children, it was said, came from Paris with bread under their arms. Now, in many cases (approximately one in ten), they arrive with the help of laboratories and with the involvement of Artificial Intelligence.
There is still no official data on how many clinics are beginning to use these technologies in their fertility treatments, but “we know that the application of these procedures can be done in all clinical areas of treatment (culture, genetics…), so “The forecast is that it will end up being just another routine in the process,” says Jordi Suñol, medical coordinator at the Instituto Bernabeu Palma de Mallorca and coordinator of the AI Unit. Not in vain, “AI improves clinical results by a percentage that should be between 12% and 15%,” as assumed at specialist conferences, also helping to collect “a very, very valuable amount of information for “so that patients can know their situation and, if the treatments do not work, be able to choose different options,” Suñol emphasizes.
A vision shared by Emilio Gómez, embryologist at the Next Fertility clinics. “If we understand success as improving productivity and efficiency in laboratory work or reducing subjectivity when making decisions, success is guaranteed,” he says.
How does it apply
The use that is being made of AI in reproductive medicine occurs in several phases of the process. «There are different applications that have been trained using embryonic morphology and kinetics, to help embryologists non-invasively choose the best embryos to transfer to the patient. There are also tools to determine oocyte quality, so that, for example, we can objectively know if the oocytes that a patient suffering from cancer has vitrified are of good quality and/or if she needs to vitrify a greater number to have more pregnancy opportunities when free of the disease,” Gómez gives as examples, who also talks about its application in the real-time recognition of sperm with optimal morphology and mobility for microinjection.
In addition to obtaining the cells, in the cultivation, fertilization and transfer, Artificial Intelligence can also be applied in the most genetic part or in the prediction of embryonic ploidy, so that it can be determined that the gametes are chromosomally normal and that do not have genetic alterations that could lead to a greater risk of abortion or malformation.
«Currently, research is being carried out with systems that are capable of determining the best hormonal treatment to stimulate the ovary of each patient, so that the treatment, doses and even the moment of inducing the maturation of the oocytes are personalized as much as possible, and even that predict whether or not the treatment will be effective,” adds Gómez.
“Thanks to all this information we manage to have better success rates, because we get patients to have progressive pregnancies as quickly as possible,” explains Suñol.
Data and ethics
As AI, the algorithms that are being used in each of the phases have been trained with a large amount of related data (embryonic oocyte quality, ovarian stimulation…). The experts in the field themselves recognize that there are patients for whom AI is more useful than others. “If we only have one embryo, we don’t need AI to tell us which one is better,” says Gómez, who adds that AI “will allow us to personalize treatments and adjust it to each patient.”
Asked about the biases that these data may contain, Suñol recognizes that it is one of the risks. “My perception is that we are in the process of beginning to understand Artificial Intelligence a little and that the recommendations it makes to us have to be evaluated with great caution,” especially taking into account that many times the answers are still pure statistics. and that it does not cover the entire population in general. “In biology, in the application of medical aspects, there are many factors that we do not even know yet and that are intervening and can condition situations that at this moment are not controlled,” Suñol contextualizes.
Therefore, “for clinicians who work with patients, communication, information and honestly transmitting what we know are much more important. It is essential that all medical care be personalized for the couple. We cannot let the statistical, mathematical and cold data of a standardized forecast of what may happen be applied to each of the couples without taking into account those particularities, personal, emotional and beliefs,” defends Gómez, while clarifying that, Although AI is an aid, both in the consultation and in the laboratory, “these systems never direct the treatments, any decision made must always be endorsed by the clinician’s experience.”
Remember that its use must always comply with ethical criteria, such as the six included in the document that the WHO published in 2021 on ethics and AI in health: -Preserve the autonomy of the patient, that is, the patient is the owner of the decisions related to their health, the privacy and confidentiality of their data must be preserved, and of course, the patient must give informed consent for any treatment they undergo.
-Promote the well-being and safety of patients and the public interest. AI systems must meet legislative and regulatory requirements for accuracy, safety and efficacy for the indications for which they are designed, and must also have quality control measures in place.
-Ensure transparency, understandability and intelligibility. Information about the AI technology being used must be published and documented. Such information must be accessible.
-Promote responsibility and accountability. Professionals must ensure that the tasks performed by AI are done under appropriate conditions and by appropriately trained people.
-Guarantee inclusion and equity. The use of AI must be designed in such a way that its use and access are as equitable as possible, without discrimination based on age, sex, gender, race, ethnicity, sexual orientation, etc.
-Promote sensitive and sustainable AI. All actors involved in the design, development and use of AI must evaluate it in real situations to determine if it adequately and appropriately responds to the needs for which it was conceived. Likewise, its environmental impact must be as small as possible.
Fewer twins and more reproductive tourism
Advances in assisted reproduction techniques are allowing for fewer high-risk pregnancies, such as multiples. According to the data from the latest registry report, there is once again a significant reduction in twinning or multiple pregnancy, placing it below 10%. «In recent years, data from the national registry have shown a progressive decrease in multiple pregnancies thanks to the increasingly widespread implementation of single embryo transfer. The single transfer is increasingly common due to technological advances and the high quality of Spanish centers, which have high pregnancy rates without the need to resort to the transfer of two or more embryos,” according to Dr. Irene Cuevas, coordinator of this registry prepared with data from the 334 public and private centers that perform fertility treatments in Spain.
Although the official registry of assisted reproduction techniques has been carried out since 2014, it is worth remembering that this report is an initiative of the Spanish Fertility Society (SEF) that started in 1993 at the request of the members of the society, to share results and thus be able to improve the success of reproductive techniques.
The advances in assisted reproduction clinics and the regulations that Spain has on the matter also mean that there is a certain reproductive “tourism” in our country. “Here we have a good tradition in medicine and, above all, a legal regulation that allows us to do many types of treatment with Artificial Intelligence according to a selection of parameters that are not acceptable in our culture and our ethics,” explains Suñol. Although these techniques can also be applied for genetic manipulation (something that, applied to “cure diseases, is fantastic”), it opens the door to uses “that are not regulated and share a risk” in their assessment.
#Testtube #children #arms