A model for malignancy probability prediction of adnexal masses

Area of Science:

• Gynecologic Oncology
• Medical Imaging
• Biostatistics

Background:

• Adnexal tumors require accurate pre-operative risk assessment.
• Distinguishing benign from malignant adnexal masses is clinically crucial.

Purpose of the Study:

• To develop a multivariate logistic regression model for predicting pre-operative malignancy probability in adnexal tumors.
• To integrate ultrasound gray-scale morphology and Doppler velocimetry data for improved diagnostic accuracy.

Main Methods:

• 117 women with adnexal masses underwent pre-operative ultrasound with Doppler analysis.
• Multivariate logistic regression analysis was applied using age, menopausal status, morphology, resistance index (RI), and pulsatility index (PI).
• Histological outcome served as the dependent variable to validate the predictive model.

Main Results:

• The final model included patient age, gray-scale morphological data, and RI as significant predictors of malignancy.
• The derived logistic regression equation allows for the calculation of malignancy probability.
• The model identified 34% malignant tumors among the 117 participants.

Conclusions:

• A novel model using multivariate logistic regression was developed to estimate adnexal tumor malignancy probability.
• This model is expected to offer superior accuracy compared to using gray-scale ultrasound or Doppler velocimetry independently.
• Further validation studies are currently underway to confirm the model's clinical utility.