In recent times, Artificial Intelligence, through its development and implementation, has demonstrated its potential to correct deficiencies present in different stages of assisted reproduction. Among them, it has helped to improve processes in the in vitro fertilization (IVF) laboratory, such as embryo selection. In this sense, from IVI we have carried out the most extensive studies on what AI can do for embryo selection.
Application of Artificial Intelligence to embryonic selection
This study, thanks to which IVI has revolutionized the embryology sector, has the largest combined casuistry in history. For this, 25,000 embryos and 4,000 patients have been analyzed, which translates into a universal, standardized and automatic embryo selection. In addition, the latest findings have been published in the North American journal Fertility and Sterility and in the European Reproductive Biology OL.
“In embryology laboratories we have applied solutions based on data, which allows us to evaluate the potential for embryo implantation. This, in turn, serves to improve the efficiency of one of the most important processes in assisted reproduction: embryo culture and selection. Thus, we have a 75% precision in the selection of chromosomally normal embryos. For its part, with the previous process through manual evaluation it is not possible to identify these embryos, regardless of the experience of the embryologist “, explains the Dr. Marcos Meseguer, embryologist and scientific supervisor of the Embryology Unit of IVI Valencia. Recently, Dr. Meseguer has been considered by Stanford University as one of the best researchers in the world, along with Professors José Remohí, Antonio Pellicer and Dr. Juan Antonio García Velasco, all of them IVI professionals.
Finding presented at the 37th edition of the ESHRE
The last work in this regard, under the title “Computer vision can distinguish between euploid and aneuploid embryos. A novel artificial intelligence (AI) approach to measure cell division activity associated with chromosomal status ” was presented yesterday at the 37th Congress of the European Society for Human Reproduction and Embryology (ESHRE). The event – one of the most important in the sector – is held online for the second consecutive year. The doctoral student Lorena Bori, from IVI Valencia, was in charge of presenting the main findings of the study, co-directed by Dr. Meseguer and Dr. Daniella Gilboa, from Tel-Aviv.
Main values presented in the ESHRE
- The technique was born with the objective of analyzing a chromosomally normal embryo –euploid– without having to apply invasive techniques, such as the extraction of cells from the blastocyst to analyze them chromosomally and know their content.
- First time that a technique based on Artificial Intelligence is able to accurately analyze the first stages of development of the embryo. In addition to quantifying the duration of cell cycles and allowing to know the diameter of the cells that form the blastocyst. Thus, an algorithm is generated that can distinguish, with 75% reliability, between a chromosomally normal or abnormal embryo.
- More than 2,500 embryos genetically analyzed at IVI Valencia, which makes up the most important casuistry in the world at a scientific level. This has made it possible to demonstrate that embryos, depending on their chromosomal content, have a different behavior and development pattern, and that this can be analyzed automatically by image analysis.
- In this sense, euploid embryos begin their development as blastocysts earlier, compared to aneuploid embryos. The fact that it takes more time for aneuploid embryos to take longer to reach the blastocyst phase is explained by the higher level of cellular activity.
- This revolutionary advance in the selection and categorization of the best embryos, at the chromosomal level, has an impact on increased pregnancy and pregnancy rates. In addition, it reduces the chances of chromosomal abnormalities, thus obtaining an objective and reliable prediction thanks to a fast and inexpensive technique.
- It is a revolution in assisted reproduction, since it would allow avoiding invasive techniques that, in part, can affect the viability of the embryo. Thus, the current results are equaled with non-invasive PGT-A, but avoiding cost and damage to the embryo. In addition, it would mean being able to automate a process that is currently produced manually and by hand.
In summary, we can affirm that it is a revolutionary technique, absolutely non-invasive and that it would improve all current embryo selection methods.
What Artificial Intelligence can do for IVF treatments
Artificial Intelligence, in its broadest sense, involves machine learning and deep learning. This applies to any program designed to solve problems, learn from experiences, and perform tasks in imitation of human behavior.
“This system classifies embryos automatically, through directed learning methods based on the experience of expert embryologists. Thus, it detects and evaluates all the development steps of the embryo and classifies its morphology. Automated embryo selection, compared to manual, is more accurate. For this reason, the evolutionary pregnancy probability is directly related to the score percentage, with which, therefore, the patient has a greater probability of success ”, comments Dr. Meseguer.
Improvements in assisted reproduction treatments thanks to AI
- Greater capacity and anticipation of identification of the potential patient, thanks to new indicators of infertility and correlations with behavior patterns, through the analysis of big data.
- Improvements in the management of patient expectations, thanks to the analysis of predictions of IVF success, considering live births.
- Higher treatment success rates, based on personalized and individualized treatment.
- Support in clinical decision making, in each step of the laboratory process, based on algorithms and computer vision to identify the majority of viable gametes and embryos.
- Definition of the reception status of the endometrium, thanks to intelligent algorithms and biomarkers that collaborate in determining the possibilities of an optimal transfer and achievement of pregnancy.
