EUVIMEDEUVIMED entdecken
Crossref

L26/P-171 Associations between artificial intelligence (AI)-based oocyte imaging scores and subsequent embryo AI scores assessed using KIDScore D5, iDAScore and CHLOE-EQ: a retrospective study

Z. Yarkiner, Z. Atayurt, G. Raad, G. Liperis, Y. Yilancilar, O. Ammar, B. Yuksel, M. Serdarogullari

Human Reproduction · 01.07.2026 · 41 (Supplement_1)

DOI: 10.1093/humrep/deag083.507 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Human Reproduction, online erschienen am 08.07.2026. Verfasst von Z. Yarkiner und 7 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Study question Is an AI-based oocyte imaging score associated with embryo AI scores following fertilisation generated by multiple AI platforms? Summary answer Higher AI-based oocyte scores were associated with significantly higher downstream embryo AI scores across multiple platforms. What is known already Several AI systems are integrated into early embryo assessment. KIDScore assesses implantation potential based on morphokinetics, CHLOE-EQ analyses time-lapse imaging to generate an AI-based embryo viability score, and iDAScore produces a score associated with implantation likelihood. However, these models are applied at the embryo stage, limiting their utility for expectation management in oocyte cryopreservation patients and early counselling on cycle prognosis. MAGENTA is a non-invasive AI tool that assigns a 0–10 score to metaphase II oocytes, with higher scores indicating greater blastocyst potential. Whether MAGENTA scores are associated with downstream embryo AI scores across platforms remains unknown. Study design, size, duration This retrospective, single-centre study analysed 2,341 fresh metaphase II oocytes from intracytoplasmic sperm injection (ICSI) cycles performed between March 2024 and December 2025. Only fresh semen samples were included; samples with concentration <5 million/ml and motility <10% were excluded. The primary variable was the AI-based oocyte imaging score (MAGENTA), compared with embryo AI scores generated by iDAScore, KIDScore D5 and CHLOE-EQ. Participants/materials, setting, methods Oocyte images were captured prior to ICSI and scored using MAGENTA, with oocytes categorised as A (7.6–10.0; high), B (5.1–7.5; medium–high), C (2.6–5.0; low–medium), or D (0.0–2.5; low). Injected oocytes were cultured until Day 5 in an EmbryoScope+. Embryo AI scores were generated using iDAScore (v1.2.0), KIDScore D5 (v.3) based on morphokinetic annotation, and CHLOE-EQ derived from time-lapse recordings. Main results and the role of chance The mean age of women was 28.54±5.73 years. For KIDScore D5, embryos from low-quality oocytes (MAGENTA group D) had a 22% lower score than those from high-quality oocytes (group A) (Exp(B)=0.78, 95% CI 0.70–0.86; p < 0.001). Embryos from medium–high-quality oocytes (group C) had a 9% lower KIDScore D5 than group A (Exp(B)=0.91, 95% CI 0.84–0.99; p = 0.026). For iDAScore, embryos from group D had a 27% lower score than group A (Exp(B)=0.73, 95% CI 0.65–0.82; p < 0.001), while embryos from group C had a 10% lower score (Exp(B)=0.90, 95% CI 0.81–0.99; p = 0.032). For CHLOE-EQ, embryos from group D showed a 27% lower score than group A (Exp(B)=0.73, 95% CI 0.63–0.86; p < 0.001), while embryos from group C had a 12% lower score (Exp(B)=0.88, 95% CI 0.77–0.99; p = 0.036). Oocytes in MAGENTA group B did not differ from group A for any embryo AI score (all p > 0.40). Limitations, reasons for caution This was a retrospective, single-centre study, limiting causal inference and generalisability. Clinical outcomes were not assessed, restricting clinical interpretation. AI scores are model-dependent, trained on specific datasets, and may not be interchangeable or capture all biological determinants of embryo competence. Wider implications of the findings These findings show that oocyte-level AI assessment is concordant with multiple downstream embryo AI platforms, supporting the importance of oocyte quality on embryo AI outputs. Upon verification, integrating oocyte AI scoring may enable earlier prognostic stratification, improve patient counselling, and support more informed interpretation of embryo AI assessments in IVF. Trial registration number No

Zitieren

APA

Yarkiner, Z., Atayurt, Z., Raad, G., Liperis, G., Yilancilar, Y., Ammar, O., Yuksel, B., & Serdarogullari, M. (2026). L26/P-171 Associations between artificial intelligence (AI)-based oocyte imaging scores and subsequent embryo AI scores assessed using KIDScore D5, iDAScore and CHLOE-EQ: a retrospective study. Human Reproduction, 41(Supplement_1). https://doi.org/10.1093/humrep/deag083.507

Vancouver

Yarkiner Z, Atayurt Z, Raad G, Liperis G, Yilancilar Y, Ammar O, et al. L26/P-171 Associations between artificial intelligence (AI)-based oocyte imaging scores and subsequent embryo AI scores assessed using KIDScore D5, iDAScore and CHLOE-EQ: a retrospective study. Human Reproduction. 2026;41(Supplement_1). doi:10.1093/humrep/deag083.507

BibTeX
@article{Yarkiner2026,
  author = {Yarkiner, Z. and Atayurt, Z. and Raad, G. and Liperis, G. and Yilancilar, Y. and Ammar, O. and Yuksel, B. and Serdarogullari, M.},
  title = {L26/P-171 Associations between artificial intelligence (AI)-based oocyte imaging scores and subsequent embryo AI scores assessed using KIDScore D5, iDAScore and CHLOE-EQ: a retrospective study},
  journal = {Human Reproduction},
  year = {2026},
  volume = {41},
  number = {Supplement_1},
  doi = {10.1093/humrep/deag083.507}
}
Kostenlos starten →