Study: AI-based test could raise IVF success rate to 85 percent

Study data presented Wednesday at the American Society for Reproductive Medicine (ASRM) suggest that when a computer equipped with artificial intelligence (AI) was shown images of hundreds of embryos, it could predict which would lead to a live birth with 85-percent accuracy, potentially bolstering the success rate of in vitro fertilisation (IVF). Study leader Nikica Zaninovic remarked that "if AI can recognise embryos that are normal and those that have abnormal chromosomes, that will result in a reduced miscarriage and stillbirth rate," adding that "this is something new and an 85-percent success rate is huge."

To train the AI, researchers used almost 700 images of five-day-old embryos, as well as time-lapse imaging, where a photograph of embryos in an incubator is taken every 10 minutes, avoiding the risks associated with an embryo being handled.

In the study, the computer analysed images of 328 embryos implanted into prospective mothers, and correctly identified 280 that would result in a live birth. Zaninovic noted that in his clinics, women under 35 with no underlying health conditions had around a 50-percent chance of having a baby using IVF and time-lapse imaging, a rate he said could increase to between 60 percent and 70 percent using AI.

The researchers indicated that the technology could be available for use within five years, potentially revolutionising fertility treatment. They added that hospitals and fertility clinics would not need to acquire the computer, but could send photos of embryos to the machine and would receive results "in minutes."

Did you like this article?