Abstract
For 40 years we have relied upon morphological assessment of the human embryo for its selection for transfer. With the advent of time-lapse microscopy (TLM), we have been able examine the timings of key developmental events and create algorithms to further advance embryo selection. However, such algorithms do not utilise all available data. Artificial intelligence (AI) can utilise the hundreds of successive images of each embryo as it develops to produce highly predictive rankings. This approach increases the speed of assessment, while decreasing variation typically associated with the subjective analysis of individual embryologists. Therefore, AI represents a means to increase both accuracy and standardisation of embryo selection. In the IVF laboratory, beyond its current function in embryo selection, AI will be used for gamete selection prior to fertilisation, and to increase laboratory function by its integration into Quality Control and Management systems. Although TLM combined with AI holds great potential, it does not quantitate the physiological status of the embryo. Analysis of spent culture medium has revealed that metabolism (glucose uptake and the utilisation of amino acids) of the human blastocyst is related to pregnancy. Currently, analysis of metabolic function by individual embryos requires highly specialised technologies, but with the advent of microfluidics and microfabrication we are entering an era of novel benchtop technologies capable of rapid single embryo analysis. Subsequently, it is envisaged that AI will be used in combination with TLM and the analysis of spent culture medium (to quantitate metabolism and plausibly cell-free DNA), to identify non-invasively the healthiest embryo for transfer.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have