Breakthrough in Endometrial Cancer Diagnosis: Development of an Ultrasound-Based Machine Learning Model

Strategic Insights -

Researchers at the Fujian Provincial Maternal and Children's Hospital in China have made a significant breakthrough in the diagnosis of endometrial cancer. A new study published in the Journal of Ultrasound in Medicine has developed and validated an ultrasonography-based machine learning model for predicting malignant endometrial and cavitary lesions. The model uses a combination of ultrasound features, including endometrial-myometrial junction, endometrial thickness, and endometrial echogenicity, to accurately predict the presence of cancer in premenopausal and postmenopausal women.

Key Takeaways:

  • The study used a retrospective approach, analyzing data from 2021 to 2023, and examined 1080 patients with pathologically confirmed results.
  • The machine learning model, developed using the XGBoost algorithm, demonstrated excellent predictive performance, particularly in postmenopausal patients, with an area under the curve (AUC) of 0.968.
  • SHapley Additive exPlanations (SHAP) analysis identified key predictors of malignancy, including endometrial-myometrial junction, endometrial thickness, and endometrial echogenicity.
  • The study demonstrated that the XGBoost-based model exhibited excellent predictive performance, with high sensitivity and specificity, particularly in postmenopausal patients.
  • The research concluded that SHAP analysis further enhances interpretability by identifying key ultrasonographic predictors of malignancy.

Statistics:

  • 1080 patients were included in the study, with 6 cases having a non-measurable endometrium.
  • 641 patients were premenopausal, and 433 were postmenopausal.
  • The area under the curve (AUC) for the premenopausal group was 0.845, with a sensitivity of 0.588 and specificity of 0.923.
  • The AUC for the postmenopausal group was 0.968, with sensitivity and specificity both being high at 0.895 and 0.931, respectively.
  • The study demonstrated that the XGBoost-based model exhibited excellent predictive performance, with high sensitivity and specificity in postmenopausal patients.

Sources:

  • "Development of a High-performance Ultrasound Prediction Model for the Diagnosis of Endometrial Cancer". Journal of Ultrasound in Medicine, 2025.
  • Zongjie Weng et al. "New Endometrial Cancer Study Results from Department of Ultrasound Described (Development of a High-performance Ultrasound Prediction Model for the Diagnosis of Endometrial Cancer)". OBGYN & Reproduction Week. October 27, 2025; p 451.