Back

A Non-Invasive Ultrasound-Based Model for Assessing Endometrial Cancer Risk in Postmenopausal Bleeding

Khatoon, N.; Mohsin, Z.; Wahab, S.; Akhtar, K.

2025-02-21 obstetrics and gynecology
10.1101/2025.02.17.25322434 medRxiv
Show abstract

AimsThis study evaluates the risk of endometrial carcinoma in postmenopausal women presenting with bleeding using an ultrasound-based risk scoring system. It aims to integrate imaging parameters and clinical variables for improved differentiation between benign and malignant endometrial conditions. MethodsA prospective analysis was conducted on postmenopausal women presenting with bleeding per vagina. Clinical variables including age, parity, BMI, hypertension, and diabetes were combined with ultrasound findings such as endometrial thickness, echogenicity, endo-myometrial interface, vascular patterns and Doppler scores. A comprehensive scoring system, based on International Endometrial Tumor Analysis (IETA) terminology, was employed to stratify patients risk levels. Diagnostic accuracy was determined by comparing the scoring system outcomes with histopathological findings. ResultsThe ultrasound-based scoring system demonstrated significant diagnostic accuracy in differentiating between benign and malignant endometrial pathologies. Endometrial thickness greater than 4 mm, heterogeneous echogenicity, irregular endo-myometrial interface and increase vascular flow were identified as strong predictors of malignancy. The model achieved a high sensitivity and specificity, providing a non-invasive yet reliable tool for risk stratification in clinical settings. The findings emphasize the potential of an integrated ultrasound-based scoring system in guiding clinical decisions for postmenopausal women with bleeding. ConclusionWe concluded that by incorporating both clinical and imaging parameters improves diagnostic accuracy, minimizes unnecessary invasive procedures, and aids in timely intervention.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.