Will artificial intelligence also change the face of Reproductive Medicine, as it is already doing in other areas of Medicine? At this year’s Eshre, the thirty-eighth appointment with the European Society of Human Reproduction and Embryology, which again sees in Milan, after two years of virtual meetings, 7000 delegates from all over the world, the use of artificial intelligence is very pop topic, protagonist of plenary sessions, twitter, and facebook direct.
An eye that is always the same
But what can artificial intelligence do? The watchword is standardization. Where the human eye may have some uncertainty, the machine eye follows the evolution of the embryos created in vitro in the PMA laboratories – through the incubator videos in time-lapse, frame by frame – to be able to identify that embryo that has a good chance of leading to pregnancy, for example. And it follows the evolution even beyond the expected 5 days, reaching up to seven, to identify a slower but still good quality embryo, which would have been discarded on the fifth day (Generalife study). And there are also those who hypothesize – data from another study presented in Eshre – that the analysis of a more detailed visualization may in the future even replace some pre-implantation genetic analyzes (study Ivi).
The choice of the center
Obviously we are talking about state-of-the-art laboratories, which can afford both the equipment to have the embryos developed in time-lapse – cost about 120 thousand euros per machine – and the computers (and machine learning) capable of following the development and producing hypotheses. Laboratories that have at least 400 cycles per year under their belt. While in Italy, however, many centers do not exceed a hundred compared to the thousands of cycles of the largest ones. Again the numbers – and the expertise behind those numbers – makes a difference in terms of results.
AI does not improve the chances of pregnancy, but it does save time
“I do not expect the evaluation of the algorithm to be better than that of the embryologist – he said Danilo Cimadomo, molecular biologist and head of Generalife Research, who presented two studies at the Eshre – but which instead helps to standardize decisions. Artificial intelligence as well as the time lapse exam does not improve the chances that the embryo will implant but would save time. “That very precious time considering the not young age of the women who approach the PMA. Italian average about 36 , 5 years. “In our centers the average is 38.5 and we are lucky if we manage to find two competent embryos – admits Cimadomo – in the USA, on the other hand, they access younger, at about 34.5 years, and those years of difference are an abyss for the success rates “.
Identify the patient’s success profile
But developing an algorithm that routinely examines embryos up to the fifth day is not easy. Also because the idea is to go further. “The goal – continues Cimadomo – is to take into account not only the development of the embryo but also the characteristics of the patient: age, first of all, but also ovarian reserve, weight (overweight or obese women have worse performance, says another study, like the anorexics), years in which a pregnancy is sought, but also characteristics of the semen “. Putting all these data together we could calculate the success of the technique after one, two or three treatments, for example. “
No delegation to the machine but on the contrary the use of the information provided to improve patient care. “I am against technology at all costs and certainly a machine will not be able to replace the precious work of embryologists – continues Cimadomo – also because artificial intelligence does not improve the pregnancy rate of the centers, but the work of the operators”.
In the laboratories advanced technologies
After all, technology has been at home for some time in the advanced laboratories of Pma. “Automation in our laboratories has been around for some time – he says Giovanni Coticchio, Italian member of the Eshre executive committee and of the committee that wanted and organized the congress in Milan – and it is essential to make the treatments more effective with processes that partly eliminate manual skill, the machines for cryopreserving oocytes for example are automatic or semi automatic, with advantages comparable to those obtained by human hand, but with the certainty of always being able to offer the exact same result. And the same goes for image analysis systems that evaluate the quality of the embryo. “A long time has passed since the birth of the first test tube baby (Louise Brown in 1978) and Reproductive Medicine is evolving rapidly, also to make facing the undeniable collapse of the fertility of our species.