Assessment and prediction models for the quantitative and qualitative reserve of the ovary using machine learning.

IF 4.2 3区医学 Q1 REPRODUCTIVE BIOLOGY

Journal of Ovarian Research Pub Date : 2025-07-18 DOI:10.1186/s13048-025-01732-0

Hiroshi Koike, Miyuki Harada, Kaname Yoshida, Katsuhiko Noda, Chihiro Tsuchida, Toshihiro Fujiwara, Akari Kusamoto, Zixin Xu, Tsurugi Tanaka, Nanoka Sakaguchi, Chisato Kunitomi, Nozomi Takahashi, Yoko Urata, Kenbun Sone, Osamu Wada-Hiraike, Yasushi Hirota, Yutaka Osuga

{"title":"Assessment and prediction models for the quantitative and qualitative reserve of the ovary using machine learning.","authors":"Hiroshi Koike, Miyuki Harada, Kaname Yoshida, Katsuhiko Noda, Chihiro Tsuchida, Toshihiro Fujiwara, Akari Kusamoto, Zixin Xu, Tsurugi Tanaka, Nanoka Sakaguchi, Chisato Kunitomi, Nozomi Takahashi, Yoko Urata, Kenbun Sone, Osamu Wada-Hiraike, Yasushi Hirota, Yutaka Osuga","doi":"10.1186/s13048-025-01732-0","DOIUrl":null,"url":null,"abstract":"Background: The age-related decline of fertility is caused by a reduction of the ovarian reserve, which is represented by the number and quality of oocytes in the ovaries. Anti-Müllerian hormone (AMH) is considered one of the most useful markers of the quantity of the ovarian reserve; however, a more accurate prediction method is required. Furthermore, there is no clinically useful tool to assess the quality of the ovarian reserve and therefore a prediction tool is required. Our aim is to produce a model for prediction of the ovarian reserve that contributes to preconception care and precision medicine.Methods: This study was a retrospective analysis of 442 patients undergoing assisted reproductive technology (ART) treatment in Japan from June 2021 to January 2023. Medical records and residual serum of patients undergoing oocyte retrieval were collected. Binary classification models predicting the ovarian reserve were created using machine learning methods developed with many collected feature values. The best-performing model among 15 examined models was selected based on its area under the receiver operating characteristic curve (AUC) and accuracy. To maximize performance, feature values used for model creation were narrowed down and extracted.Results: The best-performing model to assess the quantity of the ovarian reserve was the random forest model with an AUC of 0.9101. Five features were selected to create the model and consisted of data from only medical records. The best-performing model to assess the quality of the ovarian reserve was the random forest model, which had an AUC of 0.7983 and was created with 14 features, data from medical records and residual serum analysis.Conclusion: Our models are more accurate than currently popular methods for predicting the ovarian reserve. Furthermore, they can assess the ovarian reserve using only information obtained from a medical interview and single blood sampling. Enabling easy measurement of the ovarian reserve with this model would allow a greater number of women to engage in preconception care and facilitate the delivery of personalized medical treatment for patients undergoing infertility therapy.Clinical trial number: Not applicable.","PeriodicalId":16610,"journal":{"name":"Journal of Ovarian Research","volume":"18 1","pages":"153"},"PeriodicalIF":4.2000,"publicationDate":"2025-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12273052/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Ovarian Research","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1186/s13048-025-01732-0","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"REPRODUCTIVE BIOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Background: The age-related decline of fertility is caused by a reduction of the ovarian reserve, which is represented by the number and quality of oocytes in the ovaries. Anti-Müllerian hormone (AMH) is considered one of the most useful markers of the quantity of the ovarian reserve; however, a more accurate prediction method is required. Furthermore, there is no clinically useful tool to assess the quality of the ovarian reserve and therefore a prediction tool is required. Our aim is to produce a model for prediction of the ovarian reserve that contributes to preconception care and precision medicine.

Methods: This study was a retrospective analysis of 442 patients undergoing assisted reproductive technology (ART) treatment in Japan from June 2021 to January 2023. Medical records and residual serum of patients undergoing oocyte retrieval were collected. Binary classification models predicting the ovarian reserve were created using machine learning methods developed with many collected feature values. The best-performing model among 15 examined models was selected based on its area under the receiver operating characteristic curve (AUC) and accuracy. To maximize performance, feature values used for model creation were narrowed down and extracted.

Results: The best-performing model to assess the quantity of the ovarian reserve was the random forest model with an AUC of 0.9101. Five features were selected to create the model and consisted of data from only medical records. The best-performing model to assess the quality of the ovarian reserve was the random forest model, which had an AUC of 0.7983 and was created with 14 features, data from medical records and residual serum analysis.

Conclusion: Our models are more accurate than currently popular methods for predicting the ovarian reserve. Furthermore, they can assess the ovarian reserve using only information obtained from a medical interview and single blood sampling. Enabling easy measurement of the ovarian reserve with this model would allow a greater number of women to engage in preconception care and facilitate the delivery of personalized medical treatment for patients undergoing infertility therapy.

Clinical trial number: Not applicable.

Abstract Image

查看原文本刊更多论文

基于机器学习的卵巢定量和定性储备评估和预测模型。

背景：与年龄相关的生育能力下降是由卵巢储备减少引起的，这是由卵巢中卵母细胞的数量和质量所代表的。抗勒氏激素（AMH）被认为是卵巢储备数量最有用的标志物之一；然而，需要一种更准确的预测方法。此外，没有临床上有用的工具来评估卵巢储备的质量，因此需要一种预测工具。我们的目标是建立一个预测卵巢储备的模型，有助于孕前护理和精准医学。方法：本研究对日本2021年6月至2023年1月接受辅助生殖技术（ART）治疗的442例患者进行回顾性分析。收集取卵患者的病历及残血清。预测卵巢储备的二元分类模型使用机器学习方法开发了许多收集到的特征值。根据受试者工作特征曲线下面积（AUC）和准确度，从15个模型中选出表现最佳的模型。为了使性能最大化，用于模型创建的特征值被缩小并提取。结果：随机森林模型是评价卵巢储备数量的最佳模型，AUC为0.9101。我们选择了五个特征来创建模型，并且只包含来自医疗记录的数据。评估卵巢储备质量的最佳模型是随机森林模型，AUC为0.7983，由14个特征、医疗记录数据和剩余血清分析创建。结论：我们的模型比目前流行的预测卵巢储备的方法更准确。此外，他们可以仅使用从医学访谈和单次血液采样中获得的信息来评估卵巢储备。使用该模型可以方便地测量卵巢储备，这将使更多的妇女能够进行孕前护理，并有助于为接受不孕症治疗的患者提供个性化的医疗服务。临床试验号：不适用。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Ovarian Research REPRODUCTIVE BIOLOGY-

CiteScore

6.20

自引率

2.50%

发文量

125

审稿时长

>12 weeks

期刊介绍： Journal of Ovarian Research is an open access, peer reviewed, online journal that aims to provide a forum for high-quality basic and clinical research on ovarian function, abnormalities, and cancer. The journal focuses on research that provides new insights into ovarian functions as well as prevention and treatment of diseases afflicting the organ. Topical areas include, but are not restricted to: Ovary development, hormone secretion and regulation Follicle growth and ovulation Infertility and Polycystic ovarian syndrome Regulation of pituitary and other biological functions by ovarian hormones Ovarian cancer, its prevention, diagnosis and treatment Drug development and screening Role of stem cells in ovary development and function.