Optimise the selection of the most suitable embryo for transfer in Assisted Reproduction Cycles
Achilleas Papatheodorou, PhD, M.Med.Sc., Senior Clinical Embryologist (certified by ESHRE),
Director of the Embryolab Laboratories
It is widely and globally accepted that the fertility of couples diminishes over time. The World Health Organization (WHO) has defined infertility as a “disease” of the reproductive system. Infertile couples resort to assisted reproduction treatments to find a solution to their problem. The science of assisted reproduction is relatively new and keeps progressing year by year. Despite all the clinical, laboratory and technological advancements in this field, the effectiveness of these treatments remains low. Only 1/3 of the couples who resort to In Vitro Fertilization (IVF) will manage to have a child. There is an imperative need to find effective tools that will improve the clinical result. Essentially, it is a social need.
IVF today
The IVF process includes the implementation of a clinical protocol, which results in the couple undergoing treatment having many embryos available for transfer into the woman’s uterus. The legislation but also responsible clinical practice ideally demand transferring maximum one or two embryos into the woman’s uterus. The limitation to this number exists so as to avoid multiple gestation, which, in some cases, may actually prove dangerous for the expectant mother.
The most crucial stage in IVF treatment is the selection of the most suitable embryo for transfer. If this is done correctly, the couple will have a positive result and will not need to resort to new treatment, which alleviates the emotional, physical and financial stress.
Embryo assessment
The traditional approach to embryo assessment is performed manually. It includes the assessment of the embryos by trained scientists, known as clinical embryologists, who use light microscopes to observe specific visual specifications. At the end of the assessment, the embryologist assigns a description/score to each embryo, trying to place them in an order of preference. The embryo with the highest score is considered the most suitable for transfer. The grading system most commonly used by embryologists is the Gardner system, which checks and grades certain areas of the embryo on the 5th day of its development, using alphanumeric scores: the inner cell mass (ICM), the trophectoderm (TE) and the rate of development of the embryo’s distinctive cavity at this stage.
Unfortunately, embryo assessment carries a high degree of subjectivity and, as a result, great variations in grading are observed both between different laboratories and within the same laboratory. So it is clearly evident that it is hard to achieve standardisation in embryo assessment within one laboratory, all the more at an international level.
How can technology assist in embryo assessment?
The use of new technology in assisted reproduction attempted to provide better prospects for selecting suitable embryos, aiming to increase the effectiveness of IVF treatments. The embryos were placed in time-lapse (TL) technology incubators, meaning chambers with micro cameras, so they would be under constant observation. It was believed that this practice would assist embryologists to select the most suitable embryo and that it would be easier for the couple to eventually achieve pregnancy. Many couples were benefited by this technology, but a recent review of international literature revealed that there is room for improvement in the effectiveness of these treatments. Besides, even in this approach, the embryologists still select the embryo for transfer in a subjective manner. On many occasions, they may do so under emotional pressure. Additionally, this technology is so expensive that just 5% of laboratories globally were able to incorporate it into their day-to-day practice.
What is artificial intelligence?
“Αrtificial intelligence” (AI) is the field of computer science that designs and develops IT systems that simulate elements of human behaviour which imply at least basic intelligence: learning, adaptability, extraction of conclusions, context-based understanding, problem-solving, etc.
AI offers to machines the ability of understanding their environment, solving problems and acting towards achieving a specific goal. The computer receives data (readily available or collected via sensors, e.g. a camera), processes them and responds based on them.
The AI systems are capable of adapting their behaviour, to a certain extent, analysing the consequences of previous actions and solving problems autonomously. An impressive feat of their nature is that they learn from their mistakes and they self-improve.
AI in assisted reproduction
The latest developments in the field of AI have helped optimise medical procedures. What we are expecting in assisted reproduction in the coming years is the development of certain IT systems – “machines” – that will be able to process and analyse a multitude of data relating to a couple (e.g. demographic, hormonal, genetic, information about the couple’s laboratory attempts, etc.) and eventually link them to the videos of the embryos in pre-implantation development created using the timelapse incubator technology explained above. Then, again with the help of AI, these machines will be trained to show us which embryo is the most suitable for transfer.
This new scientific and technological approach aims to reduce to a minimum the time needed for an infertile couple to achieve pregnancy.
Use of AI in Embryolab
In the last two years, Embryolab has been participating in the development of an AI system, a non-invasive tool, that will feature automations and will indicate which embryo is most suitable for producing a much-wanted pregnancy. This technological product is the result of the partnership between Embryolab and AiVF, an Israeli tech and innovation company specialising in the development of AI systems for the field of assisted reproduction. The software this machine uses has been trained by tens of thousands of time-lapse videos with in-vitro embryo development footage and, guided by the Embryolab embryologists, has learned to perceive the stage of development the embryos are at and successfully grade them. In addition, this “machine” is able to perceive hundreds of morphokinetic events in embryo development and combine them with a positive or negative result. This is a continuous learning process that allows the system to self-improve. So, as time goes by, we are expecting that this system will develop the kind of AI technology that will allow us to identify the most suitable embryo for transfer not just based on the obvious features of an embryo, but also based on data the human eye cannot perceive. So we are aspiring for these innovative systems to increase the effectiveness of IVF treatments and bring about change in the field of assisted reproduction in general.
