Human embryo selection in in vitro fertilisation (IVF) treatments is traditionally based on evaluating cell number, morphology, and fragmentation to select the most viable embryos. However, this static evaluation method has limitations, as it captures only a snapshot of the embryo at a single time point. Recent advancements in time-lapse imaging and artificial intelligence (AI) have improved embryo assessment by enabling dynamic and continuous observation of embryonic development.This retrospective cohort study aimed to evaluate morphometric and morphokinetic parameters of 102 human embryos cultured during IVF cycles. Embryos were included if they reached the blastocyst stage on Day 5 and were selected for fresh single embryo transfer. Morphological and morphokinetic parameters were assessed using time-lapse technology to compare embryos resulting in clinical pregnancy and those that did not.Clinical pregnancy was achieved in 47.1% of transfers. Although no significant differences were observed in blastocyst area, diameter, or inner cell mass (ICM) size between embryos that implanted (P+) and those that did not (P-), trends toward larger dimensions were observed in the P+ group. Morphokinetic parameters showed slightly faster developmental kinetics in P+ embryos, although not significantly.Morphokinetic scores were calculated using time-lapse embryo evaluation systems (KIDScore D5 and iDAScore, Vitrolife). iDAScore values were significantly higher in implanted embryos. iDAScore, which incorporates AI-based scoring, showed better predictive performance compared with KIDScore in ROC and logistic regression analyses, indicating its potential as a reliable tool for embryo selection. These findings support the use of AI-based morphokinetic analysis for improved embryo selection in IVF.
Zakár et al. (Thu,) studied this question.