EUVIMEDEUVIMED entdecken
Crossref

L26/O-030 Chromosomal complexity correlates to AI-driven embryo scores across aneuploidy subtypes in PGT-A blastocysts

M. Cremades Rodelgo, J.J. Fraire-Zamora, I. Miguel-Escalada, M.J. Zamora, A. Rodriguez Aranda, K. Kelley, K. Wiemer, A. Ben-Meir, P. Pini

Human Reproduction · 01.07.2026 · 41 (Supplement_1)

DOI: 10.1093/humrep/deag083.030 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Human Reproduction, online erschienen am 08.07.2026. Verfasst von M. Cremades Rodelgo und 8 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Study question Can an AI embryo scoring model trained for predicting implantation differ across specific categories of chromosomal abnormalities? Summary answer Severe chromosomal abnormalities are associated with poorer embryo scores while chromosomal gaining abnormalities show scores comparable to euploid embryos. What is known already Early embryo development is impacted by both genetic and non-genetic factors, and not all aneuploid embryos exhibit abnormal morphokinetics and morphometrics. Different studies demonstrate that some aneuploid embryos can progress through early cleavage and even reach the blastocyst stage with morphokinetic and morphometric features comparable to euploid embryos. AI-driven embryo scores can integrate information from embryo developmental timing, expansion dynamics, and morphological changes. Prior studies demonstrate that certain embryos with chromosomal abnormalities may appear developmentally normal during early stages. However, the relationship between specific aneuploidy diagnosis and developmental dynamics remains incompletely understood. Study design, size, duration This retrospective observational study included 1,950 human blastocysts undergoing preimplantation genetic testing for aneuploidies (PGT-A). Embryos were classified as either euploid (n = 704) or aneuploid (n = 1,246), according to their PGT-A results. Aneuploid embryos were further categorized by type of anomaly: single chromosomal gains (n = 253), single chromosomal losses (n = 329), double chromosomal gains (n = 65), double chromosomal losses (n = 100), a combination of one chromosomal gain and one chromosomal loss (n = 177), and multiple (≥3) abnormalities (n = 322). Participants/materials, setting, methods Embryo quality scores (scale 1–10) were generated from timelapse sequences based on morphokinetic and morphometric features using an AI scoring model trained for predicting implantation with videos from more than 74.000 developing embryos. Beta mixed-effects regression was used to account for patient-level clustering, and discriminatory performance was evaluated using AUCs from logistic mixed-effects models. Main results and the role of chance Mean female age was 38.8 (±4.4). Mean AI-derived embryo quality scores differed across embryo chromosomal categories. Euploid embryos demonstrated the highest mean score (7.14; 95% CI: 6.90–7.37). Embryos with chromosomal gains showed scores comparable to euploid embryos including single gains (7.09; 95% CI: 6.81–7.36), double gains (7.01; 95% CI: 6.42–7.57), and combined gain-and-loss abnormalities (7.09; 95% CI: 6.77–7.39). Embryos with chromosomal losses showed progressively lower scores, including single losses (6.71; 95% CI: 6.37–7.03), double losses (6.17; 95% CI: 5.55–6.76), and embryos with multiple (≥3) abnormalities (5.99; 95% CI: 5.62–6.35). Significantly reduced scores were observed for embryos with single losses (estimated difference −2.02; p = 0.019), double losses (−4.36; p = 0.002), and multiple abnormalities (−5.13; p < 0.001). Compared to euploid embryos, the odds of achieving a high AI embryo quality score were reduced by 18.5% for embryos with single losses (OR = 0.817; 95% CI: 0.690–0.967), 35.3% for double losses (OR = 0.647; 95% CI: 0.493–0.848), and 40.1% for embryos with multiple abnormalities (OR = 0.599; 95% CI: 0.502–0.715). Discriminatory performance between categories was observed, with AUCs of 0.603, 0.623, and 0.639, respectively. Results suggest that aneuploid embryos with certain chromosomal abnormalities can be discriminated from euploid embryos by AI-driven scores. Limitations, reasons for caution The retrospective, single-center design limits causal inference. Mosaicism and chromosome-specific effects were not directly assessed and may influence developmental dynamics. Wider implications of the findings An AI model trained for implantation prediction can identify blastocysts with chromosomal losses. Findings indicate that prioritizing biopsies in embryos with higher AI scores derived from patients with increased risk of aneuploidy may increase the chances to identify euploid embryos. Trial registration number No

Zitieren

APA

Rodelgo, M. C., Fraire-Zamora, J., Miguel-Escalada, I., Zamora, M., Aranda, A. R., Kelley, K., Wiemer, K., Ben-Meir, A., & Pini, P. (2026). L26/O-030 Chromosomal complexity correlates to AI-driven embryo scores across aneuploidy subtypes in PGT-A blastocysts. Human Reproduction, 41(Supplement_1). https://doi.org/10.1093/humrep/deag083.030

Vancouver

Rodelgo MC, Fraire-Zamora J, Miguel-Escalada I, Zamora M, Aranda AR, Kelley K, et al. L26/O-030 Chromosomal complexity correlates to AI-driven embryo scores across aneuploidy subtypes in PGT-A blastocysts. Human Reproduction. 2026;41(Supplement_1). doi:10.1093/humrep/deag083.030

BibTeX
@article{Rodelgo2026,
  author = {Rodelgo, M. Cremades and Fraire-Zamora, J.J. and Miguel-Escalada, I. and Zamora, M.J. and Aranda, A. Rodriguez and Kelley, K. and Wiemer, K. and Ben-Meir, A. and Pini, P.},
  title = {L26/O-030 Chromosomal complexity correlates to AI-driven embryo scores across aneuploidy subtypes in PGT-A blastocysts},
  journal = {Human Reproduction},
  year = {2026},
  volume = {41},
  number = {Supplement_1},
  doi = {10.1093/humrep/deag083.030}
}
Kostenlos starten →