EUVIMEDEUVIMED entdecken
Crossref

L26/P-173 A noninvasive machine learning model for predicting early embryonic developmental arrest using ART data

S.I. Yoo, S.M. Jung, G.K. Lee, J.H. Shin, H.J. Go, S.B. Jung, J.M. Lee, Y.A. Seo, S.Y. Shin, J.E. Ko, K.H. Choi, J.H. Kim

Human Reproduction · 01.07.2026 · 41 (Supplement_1)

DOI: 10.1093/humrep/deag083.509 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Human Reproduction, online erschienen am 08.07.2026. Verfasst von S.I. Yoo und 11 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Study question Can a noninvasive machine learning model using routinely collected assisted reproductive technology variables accurately predict early embryonic developmental arrest? Summary answer A Random Forest model accurately predicted early embryonic developmental arrest using standard ART variables, achieving high discriminative performance across full and blastocyst-enriched datasets. What is known already Embryonic developmental arrest is a major limiting factor in assisted reproductive technology. Conventional morphology-based assessment and invasive genetic testing provide limited predictive insight, while many artificial intelligence approaches rely on costly imaging systems that are not universally accessible. Study design, size, duration This retrospective single-center study analyzed 4,507 in vitro fertilization cycles performed between January and December 2024, comparing developmental arrest and non-arrest outcomes, with additional analysis of a blastocyst-enriched subset. Participants/materials, setting, methods Clinical and laboratory variables related to assisted reproductive technology were collected and filtered using correlation analysis, retaining twenty-one features. Three decision tree–based machine learning models were trained using ten-fold cross-validation. Feature importance was evaluated using information gain. Multidimensional relationships were explored using factor analysis of mixed data and principal component analysis. Main results and the role of chance Among the evaluated models, the Random Forest classifier demonstrated the highest performance, achieving an accuracy of 89.8 percent and an area under the receiver operating characteristic curve of 0.973. Reduced-feature models showed comparable performance. Key predictors included total oocyte count, fertilization rate, anti-Müllerian hormone level, and female age. In the blastocyst-only subset, predictive performance further improved, with an accuracy of 93.9 percent and an area under the receiver operating characteristic curve of 0.986. Multivariate analyses revealed distinct dimensions driven by ovarian reserve indicators and laboratory procedural variables. All findings were statistically significant, with probability values below 0.05. Limitations, reasons for caution This study was retrospective and based on a single-center dataset, which may limit generalizability. External validation using multicenter cohorts is required. Wider implications of the findings Noninvasive machine learning models using routinely collected assisted reproductive technology data may support embryo assessment and laboratory decision-making, and can be translated into accessible web-based research prototypes. Trial registration number No

Zitieren

APA

Yoo, S., Jung, S., Lee, G., Shin, J., Go, H., Jung, S., Lee, J., Seo, Y., Shin, S., Ko, J., Choi, K., & Kim, J. (2026). L26/P-173 A noninvasive machine learning model for predicting early embryonic developmental arrest using ART data. Human Reproduction, 41(Supplement_1). https://doi.org/10.1093/humrep/deag083.509

Vancouver

Yoo S, Jung S, Lee G, Shin J, Go H, Jung S, et al. L26/P-173 A noninvasive machine learning model for predicting early embryonic developmental arrest using ART data. Human Reproduction. 2026;41(Supplement_1). doi:10.1093/humrep/deag083.509

BibTeX
@article{Yoo2026,
  author = {Yoo, S.I. and Jung, S.M. and Lee, G.K. and Shin, J.H. and Go, H.J. and Jung, S.B. and Lee, J.M. and Seo, Y.A. and Shin, S.Y. and Ko, J.E. and Choi, K.H. and Kim, J.H.},
  title = {L26/P-173 A noninvasive machine learning model for predicting early embryonic developmental arrest using ART data},
  journal = {Human Reproduction},
  year = {2026},
  volume = {41},
  number = {Supplement_1},
  doi = {10.1093/humrep/deag083.509}
}
Kostenlos starten →