EUVIMEDEUVIMED entdecken
Crossref

L26/P-116 Time-lapse spatiotemporal radiomics identifies cytoplasmic texture features predictive of live birth among preimplantation genetic testing for aneuploidy-verified euploid embryos

T. Morimoto, Y. Ohara, T. Takeuchi, T. Ishikawa, Y. Murayama, R. Haraguchi

Human Reproduction · 01.07.2026 · 41 (Supplement_1)

DOI: 10.1093/humrep/deag083.452 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Human Reproduction, online erschienen am 08.07.2026. Verfasst von T. Morimoto und 5 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Study question Can differences in cytoplasmic imaging features undetectable by conventional morphology predict live birth potential in euploid embryos assessed by preimplantation genetic testing for aneuploidy (PGT-A)? Summary answer Euploid embryos leading to live births exhibited higher cytoplasmic texture homogeneity and lower intensity fluctuations, with an area under the curve (AUC) >0.80 for prediction. What is known already Even after excluding chromosomal aneuploidy via PGT-A, approximately half of the euploid embryos fail to achieve live births. Conventional morphological grading lacks the resolution to explain this discrepancy. Radiomics is an emerging computational approach that extracts high-throughput quantitative features from medical images to identify subvisual biomarkers. Although radiomics has demonstrated utility in oncology, the application of the modality to time-lapse imaging in assisted reproductive technology remains nascent. Specifically, the relationship between spatiotemporal cytoplasmic texture features derived from time-lapse imaging and live birth outcomes in chromosomally normal embryos remains largely unexplored, requiring further validation. Study design, size, duration This retrospective study employed a Python-based time-lapse spatiotemporal radiomics framework to analyse 243 PGT-A-verified euploid blastocysts from 243 cycles in 196 patients (maternal age 38.8 ± 3.6 years) undergoing single embryo transfer between January 2022 and September 2024. Only cycles with endometrial thickness ≥7 mm under optimal hormones and good-quality blastocysts (Gardner grade ≥BB) were included. Euploidy was confirmed via PGT-A, incorporating copy number variation and single-nucleotide polymorphism analyses, with all clinical outcomes recorded. Participants/materials, setting, methods Embryos were grouped by live birth or negative outcome. Time-lapse videos were analysed, with cytoplasmic regions automatically segmented for quantitative assessment. Spatiotemporal radiomic analysis was performed on the extracted cytoplasm to derive textural features from the grey-level co-occurrence matrix (GLCM). Group comparisons were performed using the Mann–Whitney U test. Multivariate logistic regression adjusted for maternal age at oocyte retrieval, pregnancy history, transfer cycles, anti-Müllerian hormone levels (AMH), and body mass index (BMI). Main results and the role of chance Cytoplasmic texture features derived from time-lapse videos were analysed in PGT-A-confirmed euploid embryos (113 live births and 130 negative outcomes). During early blastocyst development (BL1–2), as defined by the Gardner classification, several GLCM-derived texture features differed significantly between the outcome groups. Among these, the Difference Entropy, Difference Variance and Difference Average were consistently and significantly lower in embryos resulting in live births than in those with negative outcomes (p < 0.001). These GLCM difference-based metrics reflect local grey-level variations between neighbouring pixels (complexity, dispersion, and overall magnitude), and their concordant reduction across all three features indicates smaller, more uniform intensity differences, consistent with reduced cytoplasmic heterogeneity and intensity fluctuation at this stage. Receiver operating characteristic analysis demonstrated good discriminative performance, with AUC values of 0.82 (95% confidence interval [CI]: 0.77–0.87), 0.82 (95% CI: 0.77–0.87), and 0.81 (95% CI: 0.76–0.86), respectively. Importantly, after adjusting for maternal age at oocyte retrieval, pregnancy history, number of transfer cycles, AMH, and BMI, none of these clinical factors were significantly associated with live births, whereas all three cytoplasmic radiomic features remained independently predictive. Limitations, reasons for caution This single-centre retrospective analysis may have been influenced by imaging conditions and platform-specific factors. Although the GLCM features are relatively robust, external validation, reproducibility testing, and prospective multicentre studies are required before their clinical implementation. Wider implications of the findings Time-lapse spatiotemporal radiomics revealed quantitative cytoplasmic features associated with live births in euploid embryos, offering a potential noninvasive approach to refine embryo selection and support personalised transfer strategies. Trial registration number No

Zitieren

APA

Morimoto, T., Ohara, Y., Takeuchi, T., Ishikawa, T., Murayama, Y., & Haraguchi, R. (2026). L26/P-116 Time-lapse spatiotemporal radiomics identifies cytoplasmic texture features predictive of live birth among preimplantation genetic testing for aneuploidy-verified euploid embryos. Human Reproduction, 41(Supplement_1). https://doi.org/10.1093/humrep/deag083.452

Vancouver

Morimoto T, Ohara Y, Takeuchi T, Ishikawa T, Murayama Y, Haraguchi R. L26/P-116 Time-lapse spatiotemporal radiomics identifies cytoplasmic texture features predictive of live birth among preimplantation genetic testing for aneuploidy-verified euploid embryos. Human Reproduction. 2026;41(Supplement_1). doi:10.1093/humrep/deag083.452

BibTeX
@article{Morimoto2026,
  author = {Morimoto, T. and Ohara, Y. and Takeuchi, T. and Ishikawa, T. and Murayama, Y. and Haraguchi, R.},
  title = {L26/P-116 Time-lapse spatiotemporal radiomics identifies cytoplasmic texture features predictive of live birth among preimplantation genetic testing for aneuploidy-verified euploid embryos},
  journal = {Human Reproduction},
  year = {2026},
  volume = {41},
  number = {Supplement_1},
  doi = {10.1093/humrep/deag083.452}
}
Kostenlos starten →