Top 10 Infertility Research Priorities: Embryo Selection
Artificial Intelligence for Embryo Selection is here!
Embryo Selection has been proposed as one of the top Infertility research priorities. Healthcare professionals, people with fertility problems and infertility researchers (healthcare funders, healthcare providers, healthcare regulators, research funding bodies and researchers) were brought together in an open and transparent process resulting in an article that was published in Human Reproduction in November 2020 outlining the top future infertility-related research priorities. The initial survey was completed by 388 participants from 40 countries, and 423 potential research questions were submitted. Fourteen clinical practice guidelines and 162 Cochrane systematic reviews identified a further 236 potential research questions.
The top 10 infertility research priorities for the four areas of male infertility, female and unexplained infertility, medically assisted reproduction and ethics, access and organization of care for people with fertility problems were identified. These top ten research priorities in each topic area outline the most pressing clinical needs as perceived by healthcare professionals, people with fertility problems and others, to assist research funding organizations and researchers to develop their future research agenda.
This post discusses research priority #4: What is the optimal method of embryo selection in IVF cycles?
An article was published in Human Reproduction in November 2020 outlining the top future infertility-related research priorities. The top ten ART research priorities are;
The primary goal of embryo selection is to produce a live birth from a single embryo transfer and to minimize detrimental outcomes, like implantation failure, miscarriage, or birth defects.
Noninvasive imaging (microscopy) systems for gametes and embryos is the primary grading and selection method
Common:
There are three techniques for imaging live, transparent specimens.
Phase contrast passes light through the sample, so the image is light and dark based on the DENSITY of the sample. HMC (Hofmann Modulation Contrast) and DIC are very similar- both are rendering an image based on the rate of change in the optical path- the slope. However, they do it different ways. In HMC a filter is used to amplify certain signals- in this case, the signal is light that has refracted as it passed through the sample. In DIC light is split into parallel beams and directed at the specimen, then they are further affected as they pass through the specimen, then recombined and analyzed on the other side. DIC relies on waves of light interfering (or not) with each other. The major problem with phase contrast is that it produces “halos” of light. The major problem with DIC is that orientation of the sample matters a lot (so swimming sperm- as they change their orientation- would not be good to view with it). Additionally, you must use glass coverslips/ slides or plates with the same refractive index – whereas with HMC you can use a plastic dish, PVP etc all having different refractive indices than the sample.
Rare:
Polscope. Visualize meiotic spindle placement and formation with polarized light. Now we predict the position of the spindle by positioning the 1st polar body at 12 on the clock when we do our ICSI injections. But the spindle is not always there, and we can damage the chromosome apparatus during the injection causing the second polar body to not be able to extrude from the egg, and failed fertilization. If the spindle has not formed when we inject the egg usually dies. So the method we have now works for 90% of eggs. We don’t usually have access to this specialized microscope.
MSOME and IMSI. Expensive experimental systems that require extended handling of the sperm, potential to cause more oxidative damage, lack of agreement on what marker of any indicates sperm DNA damage (vacuoles, cytoplasmic droplet)
Time lapse monitoring. Expensive and most labs do not have it. Incubation system where the embryos are placed and grown continuously and monitored. Right now, we take embryos from their little culture chamber and inspect them at set time points. Usually at fertilization, day 3, and then 5/6/7. Time lapse can record things we miss in between those checks. Positive selection criteria include the positioning of the pronuclei and a small organelle called the nucleoli inside them, number of blastomeres, the absence of multinucleation, early cleavage to the two-cell stage, and a low percentage of cell fragments in embryos.
Genetic screening selection techniques like PGT-A are peaking in several countries
New controversy: polygenic trait selection of embryos
There have been and continue to be many many controversies in genetic testing of embryos: namely, mosaic embryos and “segmental aneuploidy” embryos can self correct and make healthy pregnancies 30% of the time.
Algorithmic and artificial intelligence scoring generated by computers is gradually enhancing our selection process, these are becoming more widely used.
EMA by AiVF is a multi-module integrated platform that combines AI, computer vision, and big data to improve the success rates in patients undergoing IVF treatment. Its genetic evaluation tool can determine if the given embryo is genetically suitable for transfer by avoiding an invasive biopsy. AiVF has the largest embryo database in the world and the platform offers objective and automatic tools. One groundbreaking study from AiVF, tested the effectiveness of EMATM, AiVF’s multi-module integrated platform, which combines AI, computer vision, and big data. According to the study, EMATM detected significant differences between aneuploid and euploid embryos during the first five days of embryonic development. This was based on a retrospective study involving 2,500 embryos with PGT-A results – 1,000 euploid (genetically normal) embryos and 1,500 aneuploid (genetically abnormal) embryos. Aneuploid embryos were significantly more likely to reach each specific embryo developmental event later than euploid embryos and the time gaps between developmental milestones were also statistically longer in aneuploid embryos.
In general, lack of video imaging systems for continuous monitoring and single step embryo culture systems, as well as lack of data handling methodologies, are holding these back from widespread deployment.
The race has been on to find metabolomic markers for embryo selection
What is the embryo using for energy and growth, and what are its waste products?
The metabolomic profiling of embryo culture media has been performed through proton nuclear magnetic resonance (1H NMR). Researchers have discovered that the metabolomics profile is correlated with embryo reproductive potential. From the proton NMR spectrum, alanine, pyruvate, and glucose levels were reduced in the culture media of embryos that resulted in pregnancy. Glutamate levels were found to be higher compared to embryos that failed to implant, possibly due to its generation from α-ketoglutarate and ammonium, thereby lowering the potentially toxic ammonium to developing embryos. A sensitivity ― the ability to identify true implantations/pregnancies ― of 88.2 percent and a specificity ― the ability to correctly predict no implantations/pregnancies ― of 88.2 percent was achieved through 1H NMR
Protein markers from the embryo culture medium
Not too many have been discovered, but surely there are more. Healthy embryos will be producing characteristic proteins as they go about their work growing.
In one study by Noci et al., soluble human leukocyte antigen-G (sHLA-G) was isolated and considered as a possible protein marker of embryo reproductive potential. The presence of sHLA-G shows no correlation with embryo morphology, and the lack of sHLA-G in culture media has a negative predictive value. In another study in which sHLA-G-positive embryos were transferred, implantation and pregnancy rates were 44 percent and 75 percent, respectively, compared to 14 percent and 23 percent of transferred sHLA-G-negative embryos.
A protein biomarker that has been found to be upregulated and increased during embryo maturation into the blastocyst stage is a Day 5 secretome ― a set of proteins secreted from the cell ― resembling ubiquitin. Ubiquitin has been implicated in the turnover of key signaling molecules during implantation
Other non-embryonic genomic markers (cumulus cells) have potential
The cumulus cells (CCs) that surround the oocyte from fertilization until implantation have been analyzed and gene-profiled to gauge embryo potential: the likelihood of an embryo to implant and lead to a successful pregnancy. Several genes expressed in CCs have been correlated with predicting pregnancy, including cyclooxygenase 2 (COX2), steroidogenic acute regulatory protein (STAR), and pentraxin 3 (PTX3). Two upregulated biomarkers have been identified in the CCs of successful pregnancies, BCL2L11 and PCK1, which are involved in apoptosis of abnormal cells and gluconeogenesis.
