Embryo selection algorithms, including manual morphological grading and AI-based algorithms, favour XY embryos over XX embryos

• Published in: Reproductive biomedicine online Vol. 48; p. 104054
• Main Authors: Popa, T., He, C., Lau, M., Hickman, C., Ottolini, C., O'neill, H.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.05.2024
• Subjects: Obstetrics and Gynecology
• ISSN: 1472-6483
• DOI: 10.1016/j.rbmo.2024.104054

Manual embryo selection (based on blastocyst morphology) and AI-based methods (based on embryo morphokinetic parameters such as tPNf, t2, t3, t4, t5, t8, tB etc.) are widely adopted methods for embryo selection and assessment. However, all current embryo grading models assume identical embryo morphology and morphokinetics irrespective of embryo sex. Currently, the presence of any differences between XX and XY embryos has been an area of active debate. If sex differences exist at the preimplantation stage, they might influence the grades assigned by current embryo selection methods, potentially introducing bias. A retrospective study was conducted on 1411 embryos with known sex information following PGT-A, at a single centre between 2018-202, making this the largest study to date interrogating morphological and morphokinetic sex differences using time-lapse and PGT-A data. We compared the scores/grades assigned to XX vs XY embryos by three commonly used embryo selection models: manual morphological grading (Gardener system), the KIDScore (VitroLife) and CHLOE (Fairtility). These algorithms provide a representative slice of the current embryo grading landscape, encompassing manual selection, traditional machine learning and modern deep learning approaches, respectively. XY embryos were more likely than XX embryos to score higher under standard morphological grading (p<0.00001) and the KIDScore D3(p=0.0182). Overall, the trophectoderm of XY embryos was deemed of ‘higher quality’ than that of XX embryos by both manual and AI-based (KIDScore) embryo assessment models. However, no difference between XX and XY embryos was identified in CHLOE EQ scores(p=0.208). We identified that manual morphological grading and AI-based embryo selection models favour XY embryos. Our findings raise significant questions surrounding the potential for sex bias in commonly used embryo selection methods used by IVF centres worldwide. Considering the increasing reliance on assisted reproduction, any potential source of bias should be addressed to ensure equitable and ethical treatment of all embryos.