EUVIMEDEUVIMED entdecken
Studienprotokoll

L26/P-764 An Artificial Intelligence Approach to Optimizing Gonadotropin Starting Dose in GnRH-Antagonist Protocol for IVF/ICSI

M. Jing, R. Zhang

Human Reproduction · 01.07.2026 · 41 (Supplement_1)

DOI: 10.1093/humrep/deag083.1091 ↗

Kurz gefasst

Studienprotokoll in der Zeitschrift Human Reproduction, online erschienen am 08.07.2026. Verfasst von M. Jing und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Study question To develop an artificial intelligence model to predict the optimal starting dose of gonadotropins in the GnRH-antagonist protocol, using real-world clinical data from IVF/ICSI patients. Summary answer The application of artificial intelligence in this study has achieved an 84% accuracy rate in predicting retrieved optimal oocyte numbers within a ± 2-3 range. What is known already Controlled ovarian stimulation (COS) is crucial for the success of in vitro fertilization (IVF), as it allows for the retrieval of multiple oocytes. However, there is currently no standardized method for determining the optimal starting dose of gonadotropins to achieve an optimal number of oocytes, particularly in the GnRH-antagonist protocol. Clinicians typically rely on empirical experience, which may lead to suboptimal outcomes. Artificial intelligence offers a promising approach to personalize the starting dose based on patient-specific data, potentially improving the precision of ovarian stimulation. Study design, size, duration We collected clinical data from infertile patients undergoing the GnRH-antagonist protocol at the hospital. The dataset included patient age, height, weight, anti-Müllerian hormone (AMH) levels, antral follicle count (AFC), basal follicle-stimulating hormone (b-FSH), basal luteinizing hormone (b-LH), basal estradiol (b-E2), and the number of oocytes retrieved. An Extreme Gradient Boosting Model was trained to predict the optimal starting dose of gonadotropins, aiming for an ideal oocyte yield of 8-14 oocytes. Participants/materials, setting, methods Infertile patients undergoing the GnRH-antagonist protocol Main results and the role of chance The artificial intelligence model achieved an 84% prediction accuracy, where the absolute difference between the predicted ovulation number and the actual value ranged between 2 and 3.Results Data validation is ongoing. Limitations, reasons for caution The clinical data utilized were sourced from a single center, which may limit the generalizability of the findings. Moreover, live birth rate was not included as an outcome measure in this study. Wider implications of the findings The integration of big data analytics with artificial intelligence algorithms not only improves the accuracy of gonadotropin starting doses but also paves the way for more standardized and effective IVF treatment strategies. Trial registration number No

Zitieren

APA

Jing, M., & Zhang, R. (2026). L26/P-764 An Artificial Intelligence Approach to Optimizing Gonadotropin Starting Dose in GnRH-Antagonist Protocol for IVF/ICSI. Human Reproduction, 41(Supplement_1). https://doi.org/10.1093/humrep/deag083.1091

Vancouver

Jing M, Zhang R. L26/P-764 An Artificial Intelligence Approach to Optimizing Gonadotropin Starting Dose in GnRH-Antagonist Protocol for IVF/ICSI. Human Reproduction. 2026;41(Supplement_1). doi:10.1093/humrep/deag083.1091

BibTeX
@article{Jing2026,
  author = {Jing, M. and Zhang, R.},
  title = {L26/P-764 An Artificial Intelligence Approach to Optimizing Gonadotropin Starting Dose in GnRH-Antagonist Protocol for IVF/ICSI},
  journal = {Human Reproduction},
  year = {2026},
  volume = {41},
  number = {Supplement_1},
  doi = {10.1093/humrep/deag083.1091}
}
Kostenlos starten →