Artificial intelligence-driven analysis of embryo morphokinetics in singleton, twin, and triplet pregnancies

In recent years, the transfer of more than one embryo has become less frequent to diminish multiple pregnancies. Even so, there is still a risk of one embryo splitting into two or even three. This report presents the case of a triamniotic monochorionic gestation in a 35-year-old woman, obtained after the transfer of a single day 5 embryo that had been previously hatched with a laser and subsequently transferred in a fresh IVF cycle. The morphokinetics of this embryo as well as another eight that produced monozygotic twins were compared with nine embryos that achieved singleton gestations. An artificial intelligence (AI) system was used to perform this analysis objectively. Embryo divisions leading to multiple gestations were slower across all analysed parameters. This was most evident in the time taken to reach the 5-cell stage (t5) (48.74 ± 3.21 versus 44.69 ± 5.61 h) and the morula stage (tM) (85.08 ± 10.00 versus 78.13 ± 5.35 h). Moreover, blastocyst diameter (161.80 ± 26.50 versus 179.10 ± 25.88 μm) and area (21 684 ± 7267.4 versus 26 516 ± 7254.8 μm2) were smaller in these embryos, but the differences did not reach statistical significance. This is the first report, to our knowledge, to use AI to analyse and compare morphokinetic parameters between singleton pregnancies and a group of embryos that produced twin and triplet pregnancies.