Photo/Illutration Embryo culture specialists in Kobe prepare a culture solution to grow fertilized eggs. (Asahi Shimbun file photo)

Scientists have developed new image-analysis technology that uses AI to examine the quality of fertilized eggs, a development that could lead to improving in vitro fertilization pregnancy rates.

The technology can currently be used to produce and analyze high-resolution images of the growth stages of fertilized eggs of mice and other species.

It could one day enable medical workers to make more objective assessments on human egg quality, a factor related to infertility, and could help improve in vitro fertilization pregnancy rates, according to a team of researchers from Keio University, Kindai University and other institutions.

In vitro fertilization is an infertility treatment where sperm and eggs are taken from a couple for fertilization and then put back in the woman’s uterus.

Egg quality can affect pregnancy rates and chances of successful childbirth.

But in most cases, there are currently no objective criteria for medical professionals on which eggs to select. They are left to visually inspect and choose which fertilized eggs should be used for implantation based on the criteria of each individual medical institution.

The lack of standardization across the health system for the process poses some problems. It means that pregnancy rates differ between hospitals, for instance. Setting appropriate fees for the treatment is also difficult, since it is not covered under the national health insurance program and fees vary across medical institutions.

The researchers had AI software analyze about 7,500 images of the fertilized eggs of mice, roundworms and fruit flies, and developed technology that can produce 3-D images of fertilized eggs’ cell nuclei.

The nuclei are colored to make them easier to see.

The new technology can determine the speed at which fertilized eggs grow, since the number of cell nuclei increases as the cell divides.

The team said it has started to consider whether the technology can be applied for human use.

“The 3-D images produced by the AI enabled us to see the cell nuclei overlapping,” said Akira Funahashi, an associate professor of systems biology at Keio University and a member of the team. “We want to put the technology into practical use for the treatment.”

According to the researchers, the quality of eggs has a bigger impact on the growth of fertilized eggs in their early stage than sperm quality.

If the AI can properly assess the characteristics of early fertilized eggs that have a high pregnancy rate, that could pave the way for assessing which eggs are most suitable for in vitro fertilization.

The paper was published in October in the online journal Nature Research: (https://www.nature.com/articles/s41540-020-00152-8)

Related News