#361 : What Does My Embryo Score Mean? The Relationship Between KIDScore, Life Whisperer AI Score, and Embryo Euploidy

Abstract

Background and Aims: Many models have been developed for predicting an embryo’s euploidy or implantation potential. These models assign continuous scores to embryos, based on annotation performed by embryologists according to morphokinetic parameters, or based on deep learning analysis of embryo images or video, with higher scores indicating higher probability of implantation and/or euploidy. However, the actual probability of implantation or euploidy at a given score is not always clear. We aimed to investigate the relationship between scores assigned by an annotation model (KIDScoreD5 v3.1, Vitrolife), an image analysis model (Life Whisperer, Presagen), and embryo euploidy. Method: From January 2019 to May 2021, PGT-A by NGS performed on blastocysts in Thomson Hospital Kota Damansara in Malaysia identified 273 euploid and 212 aneuploid blastocysts. Since blastocyst culture in a timelapse incubator (Embryoscope+, Vitrolife) is routine in our centre, the KIDScore for each blastocyst was known. Still image data of the same embryos was uploaded to the Life Whisperer (LW) platform. Embryos were grouped based on their KIDScore and LW score, and the percentage of euploid and aneuploid embryos in each group was calculated. Results: In both models, there were more embryos in higher score groups, as only embryos with good morphology underwent PGT-A (Figures 1 and 2). In KIDScore groups up to score 8, approximately 50% of embryos were euploid, while in KIDScore groups >8-9 and >9-10, 76.6% and 75.5% of embryos were euploid respectively (Figure 1). In LW groups up to score 7, approximately 50% of embryos were euploid, while in LW groups >7-8, >8-9, and >9-10, the percentage increased to 56.1%, 58.6%, and 66.7% respectively (Figure 2). Conclusion: Increasing KIDScores and LW scores correspond with increased percentages of euploid embryos. However, the increase is not constant, especially for KIDScores. These results can be used to estimate the probability of euploidy at a given embryo score.