Related Experiment Video
Updated: Apr 28, 2026
12:42
Heterotypic Three-dimensional In Vitro Modeling of Stromal-Epithelial Interactions During Ovarian Cancer Initiation and Progression
Published on: August 28, 2012
14.8K
Age-Stratified Modeling of Clinical Heterogeneity in Polycystic Ovary Morphology Using Ultrasound-Based Machine
Chenke Kuang1,2,3,4, Qiao Wei2,3,5,6, Zichao Liu2,3,4,6
1Department of Ultrasound, Department of Medical Imaging, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China, Changsha, Hunan, People's Republic of China.
International Journal of Women'S Health
|April 27, 2026
Summary
Machine learning accurately stratified patients with polycystic ovary morphology (PCOM) by age. Diagnostic factors like ovarian volume and menstrual phase varied by age group, enabling personalized PCOM assessment.
Area of Science:
- Reproductive endocrinology and medical imaging.
- Application of artificial intelligence in healthcare.
Background:
- Polycystic ovary morphology (PCOM) diagnosis relies on clinical and ultrasound data.
- Current diagnostic approaches may not fully account for age-specific variations in PCOM indicators.
Purpose of the Study:
- To develop an automated age stratification model for PCOM patients using machine learning.
- To identify age-specific clinical and ultrasound characteristics contributing to PCOM diagnosis.
- To enhance diagnostic accuracy through personalized assessment.
Main Methods:
- Analysis of clinical and ultrasound data from 192 ovaries.
- K-means clustering for automated age stratification.
- XGBoost, Random Forest, SVM, and ANN algorithms for feature analysis.
- Five-fold cross-validation to evaluate model performance (accuracy, sensitivity, F1-score, AUC).
Main Results:
- PCOM patients were successfully stratified into three distinct age groups.
- Ovarian volume and follicle count positively correlated with age strata.
- Key diagnostic drivers differed by age: menstrual phase (18-22 yrs), pregnancies and stromal artery RI (23-29 yrs), S/D ratio and live births (30-40 yrs).
Conclusions:
- Age-specific indicators significantly influence ultrasound-based PCOM diagnosis.
- The study provides a framework for personalized diagnostic assessment in PCOM.
- Machine learning facilitates improved diagnostic accuracy by considering age-related factors.
