0136 GMT February 16, 2019
A machine has been developed that can select embryos that are likely to give a live birth with astounding accuracy of 85 percent, archyworldys.com wrote.
Researchers say robotic learning will ‘revolutionize’ IVF and become available to NHS hospitals within five years.
Couples desperately seeking children could dramatically improve birth rates as British physicians email a photograph of their embryo and receive a response within minutes.
The technology was developed jointly by Imperial College London and Cornell University in New York.
It uses time lapse photos of embryos in an incubator that currently help medics most likely to lead a baby.
About half of the early pregnancy losses are due to the fact that the embryo has an abnormal number of chromosomes.
The breakthrough was honored this week with an award at Denver, the world's largest fertility conference.
Dr. Nikica Zaninovic, who led the study at Cornell University, said, "If AI can detect normal and chromosomally abnormal embryos, it results in a reduced miscarriage and stillbirth rate, which is really our main reason for doing research.
"This research is something really new and a success rate of 85 percent is enormous."
For women under the age of 35, with no further health problems, the IVF live birth rate could increase to as much as 70 percent, said Drs. Zaninovic.
Currently, the proportion of NHS among under-35s is 30 percent.
The use of time-lapse photography technique to choose embryos without the help of AI, has been around since 2010.
The embryos remain in an incubator and a photograph is taken every 10 minutes to measure growth.
Computers then produce data to help medics choose an embryo to implant the mother.
Scientists have fed a New York supercomputer — nicknamed ‘The Beast’ — with thousands of historical images to ‘teach’ what to look for in successful embryos.
It was told those who led to a live birth in which pregnancies failed or led to stillbirths.
Dr. Zaninovic, whose team is currently patenting the technology, said, "All I need is patient information from a hospital in London and the image of the embryo, and I can put this into the computer.
"It's all web-based, it does not mean every hospital needs it, it can easily be done over the Internet."
He added, "Within five years, it will be routinely used in clinical settings."
The beast received more than 50,000 images from more than 10,000 embryos.
In a retrospective experiment, single images of 328 embryos were implanted in potential mothers.
When asked which of these would have led to a live birth, he chose 280 correctly.
The rate of 85 percent is well above what researchers of a human embryologist expect, with current standards varying from clinic to clinic.
Researchers say it is ‘almost 100 percent’ because about 15 percent of the reasons why live births do not happen are due to problems in the womb.
The findings were presented at the annual meeting of the American Society for Reproductive Medicine (ASRM).
Some UK clinics use time-lapse photography technology to help clinicians decide which embryo to choose.
Evidence of whether it is improving successful birth rates in its present form is mixed.
Some believe that these measurements help to improve the success rate, partly because embryos do not have to be handled by scientists for five days.
Leading UK fertility specialist Proffessor Allan Pacey of Sheffield University said: "The ability to observe the development of embryos using time-lapse systems has greatly developed in recent years and these devices have become commonplace in clinical practice.
"So far, however, the data has not found out that they actually help to select the best embryo and improve the chances of getting pregnant.
"Therefore, the application of artificial intelligence to data collected from time-lapse systems is a very good idea because it can find patterns and algorithms that are invisible to the human eye."
Professor Charles Kingsland founded Britain's largest NHS IVF unit, the Hewitt IVF Center in Liverpool, which was the first company to successfully use time-lapse technology.