Shear wave elastography values in endometrioma: Clinical findings and machine learning-based prediction models

Area of Science:

• Gynecologic imaging
• Medical diagnostics
• Computational pathology

Background:

• Endometriomas are common gynecological conditions often associated with significant clinical symptoms.
• Accurate non-invasive diagnostic methods are crucial for timely management.
• Shear wave elastography (SWE) offers a quantitative assessment of tissue stiffness.

Purpose of the Study:

• To assess the diagnostic value of SWE in endometriomas.
• To correlate SWE-derived stiffness with clinical symptoms such as dysmenorrhea, dyspareunia, and infertility.
• To evaluate machine learning (ML) models for predicting clinical outcomes based on SWE data.

Main Methods:

• Prospective study of 94 women (20-35 years) with unilateral ovarian endometriomas.
• SWE used to measure mean shear wave velocity (SWV) in m/s.
• Clinical symptoms evaluated; ML models (Logistic Regression, Random Forest, Gradient Boosting, SVM) applied to predict outcomes.

Main Results:

• Higher SWV observed in patients with dysmenorrhea (P=0.024), dyspareunia (P=0.016), and infertility (P=0.014).
• ML models achieved high predictive accuracy for dysmenorrhea (ROC-AUC 0.94) and dyspareunia (ROC-AUC 0.98).
• Logistic Regression and SVM demonstrated superior performance among the ML algorithms.

Conclusions:

• SWE combined with ML provides a non-invasive and cost-effective method for endometrioma diagnosis and clinical outcome prediction.
• This integration of imaging and computational tools advances precision medicine in gynecology.