EUVIMEDEUVIMED entdecken
Crossref

L26/P-150 Multivariable analysis and predictive model development for usable blastocyst formation in patients with a single oocyte undergoing ICSI: an ambispective observational study

K. Rodak, K. Trzęsicka, M. Krajewski, J. Liss

Human Reproduction · 01.07.2026 · 41 (Supplement_1)

DOI: 10.1093/humrep/deag083.486 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Human Reproduction, online erschienen am 08.07.2026. Verfasst von K. Rodak und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Study question Which factors influence blastocyst formation in patients with a single oocyte undergoing ICSI, and can they be used to predict usable blastocyst development? Summary answer Female age, progesterone (PRG) level on the last stimulation day (DS), numbers of 2PN/1PN independently influence blastocyst development and allow construction of reliable predictive model. What is known already Clinical, laboratory, and embryological factors, including characteristics such as maternal age, hormone levels, sperm parameters, and laboratory-specific conditions, have been investigated for their impact on blastocyst formation. Blastocyst development is a critical indicator of implantation potential. However, identifying independent predictors remains challenging, particularly in patients with a limited number of oocytes available for fertilization. This challenge extends to predictive modeling, as previous studies have primarily focused on patients with higher oocyte yields, and the resulting models may not apply to those with fewer oocytes. Predictive models are increasingly developed, yet comprehensive approaches are still needed for individualized outcome estimation. Study design, size, duration An ambispective analysis of 850 single-oocyte ICSI cycles (2018–2025) aimed to identify predictors of usable blastocyst formation and develop a predictive model. Variables included partners’ age, body mass index (BMI), female hormone levels (estradiol (E2), luteinizing hormone (LH), PRG on first/last DS, PRG on pick-up day), anti-Müllerian hormone (AMH), stimulation duration, predicted follicles, retrieved cumulus complexes/oocytes, oocyte assessment, sperm concentration, viability, progressive motility, laboratory timing, adjunct procedures, number of pronuclei (day 1). Participants/materials, setting, methods The analysis included fresh ICSI cycles with single oocyte qualified for fertilization and semen provided on the procedure day. Univariate logistic regression was performed to identify candidate variables (p<0.05), which were assessed for intercorrelations/multicollinearity prior to multivariable logistic regression. Independent predictors were selected using backward stepwise regression, and the predictive model was constructed from the resulting coefficients. Internal validation was performed using V-fold cross-validation, and model discrimination was evaluated with receiver operating characteristic (ROC) curves. Main results and the role of chance In univariate analysis of 34 variables, 8 were significantly associated (p<0.05) with blastocyst formation: female and male age, PRG on the last DS, AMH level in the last examination, progressive sperm motility, numbers of 2PN, 1PN, and 0PN on day 1. Due to high correlations (|R|>0.6), male age and 0PN were excluded from the multivariable logistic regression. Using backward stepwise logistic regression, four independent predictors of blastocyst formation were identified: female age (OR: 0.908, 95% CI: 0.872–0.945, p<0.001), PRG level on the last DS (OR: 0.974, 95% CI: 0.948–1.000, p=0.046), number of 2PN (OR: 40.578, 95% CI: 18.673–88.180, p<0.001), and number of 1PN on day 1 (OR: 7.356, 95% CI: 1.972–27.448, p=0.003). The predictive model was constructed using regression coefficients: logit(p) = − 0.014 − 0.096 × (female age) − 0.027 × (PRG on the last DS) + 3.703 × (number of 2PN on day 1) + 1.996 × (number of 1PN on day 1). All data comprised the training set. The area under the ROC curve (AUC) for the training set was 0.804 (95% CI 0.776–0.833; sensitivity 91%, specificity 60%). Internal validation using V-fold cross-validation yielded AUC of 0.800. The model exhibited useful discrimination. Limitations, reasons for caution Although the model shows potential clinical utility, it lacks external validation and has not been implemented in practice. Single-center data limits generalizability. Despite considering numerous variables, future studies should include additional clinical and embryological factors to further refine and enhance the predictive model. Wider implications of the findings The present study highlights that female age, PRG on the last DS, and numbers of 2PN/1PN independently predict usable blastocyst formation in single-oocyte fresh ICSI cycles. The predictive model may support personalized counseling, optimize management of low-responder patients, and guide timely interventions for planned procedures. Trial registration number No

Zitieren

APA

Rodak, K., Trzęsicka, K., Krajewski, M., & Liss, J. (2026). L26/P-150 Multivariable analysis and predictive model development for usable blastocyst formation in patients with a single oocyte undergoing ICSI: an ambispective observational study. Human Reproduction, 41(Supplement_1). https://doi.org/10.1093/humrep/deag083.486

Vancouver

Rodak K, Trzęsicka K, Krajewski M, Liss J. L26/P-150 Multivariable analysis and predictive model development for usable blastocyst formation in patients with a single oocyte undergoing ICSI: an ambispective observational study. Human Reproduction. 2026;41(Supplement_1). doi:10.1093/humrep/deag083.486

BibTeX
@article{Rodak2026,
  author = {Rodak, K. and Trzęsicka, K. and Krajewski, M. and Liss, J.},
  title = {L26/P-150 Multivariable analysis and predictive model development for usable blastocyst formation in patients with a single oocyte undergoing ICSI: an ambispective observational study},
  journal = {Human Reproduction},
  year = {2026},
  volume = {41},
  number = {Supplement_1},
  doi = {10.1093/humrep/deag083.486}
}
Kostenlos starten →