In an IVF cycle, embryologists are required to select the 'best' embryo for an embryo transfer back to the mother's womb. Embryologists have always assessed embryos by eye under a microscope in the laboratory. Most admit that, despite experience, they cannot tell accurately which embryos will develop after transfer. Arguably, the most significant advance over the last decade or more in human IVF is the adoption of technologies that attempt to determine the ‘good’ embryos that will deliver healthy babies from those that won't implant, or worse, implant but then miscarriage. IVF has seen wide adoption of two techniques that address the selection of embryos, time-lapse technology (a movie is made of the developing embryo during embryo culture) for a decade. Analysing the way in which an embryo grows was purported to help select the embryo that should be the first transferred, and those remaining for cryopreservation. However, although the idea is attractive, it failed to provide more accuracy than an experienced embryologist did. Most recent has been the combination of ‘artificial intelligence’ (AI) to select the best embryo(s) using images collected from either time-lapse movies, or the photographs of embryos taken immediately before embryo transfer. This has led a new explosion of papers and abstracts that suggests there is at least equalisation, if not better results from AI compared to the embryologist.
Predicting pregnancy establishment does not necessarily predict if the embryo will also become a healthy baby. Artificial intelligence and other new microscopic techniques are developing to achieve this.