Related Experiment Video
Updated: May 21, 2025
09:44
Evaluation of Hepatic Glucose Production in a Polycystic Ovary Syndrome Mouse Model
Published on: March 5, 2022
2.9K
A diagnostic model for polycystic ovary syndrome based on machine learning
Cheng Tong1,2, Yue Wu1, Zhenchao Zhuang3
1The First Affiliated Hospital of Zhejiang Chinese Medical University (Zhejiang Provincial Hospital of Chinese Medicine), Hangzhou, 310006, Zhejiang, China.
Scientific Reports
|March 22, 2025
Summary
This study developed a machine learning model combining anti-Müllerian hormone (AMH) and other hormones to diagnose polycystic ovary syndrome (PCOS). The logistic model showed promising diagnostic potential, highlighting AMH as a key indicator for PCOS.
Area of Science:
- Endocrinology
- Medical Diagnostics
- Machine Learning in Healthcare
Background:
- Diagnosis of polycystic ovary syndrome (PCOS) presents significant challenges in clinical practice.
- Existing diagnostic methods may benefit from enhanced auxiliary tools for improved accuracy.
Purpose of the Study:
- To develop and validate a diagnostic model for PCOS by integrating anti-Müllerian hormone (AMH) with steroid hormones and oestrogens.
- To provide additional diagnostic support and improve the accuracy of PCOS identification.
Main Methods:
- Collected serum samples from 84 PCOS patients and 75 healthy controls.
- Utilized machine learning algorithms to construct a diagnostic model using selected hormonal variables (LH, LH/FSH, E2, PRL, T, AMH, AD, COR).
- Evaluated model performance using a validation set, achieving an Area Under the Curve (AUC) of 0.86 with the logistic classification model.
Main Results:
- Identified 8 key hormonal variables, with AMH demonstrating the highest diagnostic potential.
- Successfully constructed five machine learning models, with the logistic classification model exhibiting the best overall performance.
- The developed model achieved an AUC of 0.86 on the validation set.
Conclusions:
- A machine learning-based diagnostic model combining AMH with other hormones shows significant potential for PCOS diagnosis.
- The logistic classification model performed best, underscoring the utility of this approach.
- Further validation with larger, multi-center datasets is recommended to confirm and refine the model's generalizability.
