Annonse
Became pregnant after 19 years with help from artificial intelligence
After 19 years of trying, a couple in the US is expecting their first child thanks to AI. Norway is also developing AI systems to help those struggling to conceive.
Doctors at Columbia University's fertility centre in the United States have developed an imaging technology, Sperm Tracking and Recovery (STAR), which they believe could help couples who struggle to conceive. They use artificial intelligence (AI), writes Columbia Spectator.
In an article in The Lancet, Dr. Zev Williams, who leads the team, explains that the STAR system is an AI-based microfluidic platform that can identify and isolate a few sperm cells in semen samples rapidly and in real time.
Williams describes using this method to help a couple where the man has azoospermia, a condition where no sperm cells are present in the semen.
About 1 per cent of all men and roughly 10 to 15 per cent of men with fertility issues have this condition.
Like "a needle hidden within a thousand haystacks"
Annonse
In a typical sperm sample, there may be between 100 and 300 million sperm cells. But in some cases, there are only a few. Even the best fertility experts can struggle to find them.
But the STAR system can go through several million images in just a few hours and find the few sperm cells that might be present.
Williams explained it a bit more simply to Time Magazine this summer: STAR is trained to find those "really, really, really rare sperm cells." He compares it to finding a needle hidden in a thousand haystacks. The sperm cells can be carefully extracted and used in attempts to fertilise eggs.
Pregnant for the first time
After trying a variety of different methods for 19 years and going through 15 unsuccessful IVF attempts, a childless couple became pregnant in March after using the STAR system.
The system examined a 3.5-millilitre semen sample that traditional methods had already determined to contain no sperm cells. STAR analysed 2.5 million images of the sample in about two hours and detected seven sperm cells: two moving and five unmoving.
The two moving sperm cells were isolated and injected into two mature egg cells, both of which developed into embryos. Both embryos were transferred on day three, and 13 days later, the woman tested positive on a pregnancy test for the first time in her life.
Eight weeks into the pregnancy, an ultrasound confirmed normal foetal development and a heartbeat of 172 beats per minute.
A milestone
The pregnancy was made public this summer, and the baby is due in December.
Williams calls the successful fertilisation project a milestone for AI technology in treating infertility.
"Male factors account for about 40 per cent of infertility in the US, and azoospermia is responsible for about 10 per cent of those cases. Until recently, there was little doctors could do to address the lack of sperm needed to fertilise an egg, other than using donor sperm," Williams explains.
The technology was recently named by TIME Magazine as one of the best inventions in 2025.
Still a few years away in Norway
Annonse
Norwegian researchers have also been exploring AI in fertility treatment, though not specifically for azoospermia.
Researchers from UiT the Arctic University of Norway and the University Hospital of North Norway (UNN) have created a tool called Spermotile. It uses AI to analyse individual sperm cells in semen samples to ensure high-quality selection for assisted reproduction.
"We're still in the research and development phase and aim to launch a company to commercialise the technology in 2027," business developer Lars Sørensen at Norinnova tells NTB.
Norinnova is in charge of bringing the project to market.
"The news about the STAR system is encouraging, because AI can strengthen infertility treatments," says Sørensen.
Spermotile is not designed specifically for azoospermia, but it could still support fertility clinics in dealing with it, as well as help improve and simplify fertility treatments overall.
Rapid progress worldwide
Researchers at OsloMet have been studying AI-assisted fertility treatment for years, primarily focusing on identifying the best sperm cells and embryos.
"We were early internationally, but the global development in this field has exploded," says Trine B. Haugen, professor emeritus of biomedicine at OsloMet.
In the American project, the goal was to find sperm that cannot be detected through traditional methods and isolate any that were moving.
"For men with azoospermia, one way to obtain sperm has been to extract cells from the testicle. But this often fails and requires surgery. The American approach may be a better option for this patient group," says Haugen.
Currently, OsloMet’s work is focused on improving AI for analysing semen samples faster and more effectively than the methods used today in infertility evaluations.
———
Translated by Alette Bjordal Gjellesvik
Read the Norwegian version of this article on forskning.no
Related content:
Subscribe to our newsletter
The latest news from Science Norway, sent twice a week and completely free.
