A multivariate logistic regression analysis in predicting malignancy for patients with ovarian tumors

Area of Science:

• Gynecologic Oncology
• Medical Imaging
• Diagnostic Ultrasound

Background:

• Accurate preoperative diagnosis of ovarian malignancy is crucial for effective treatment planning.
• Gray-scale ultrasonography has limitations in differentiating benign from malignant ovarian tumors.
• Doppler ultrasound variables offer potential for improved diagnostic accuracy.

Purpose of the Study:

• To enhance the preoperative diagnosis of ovarian malignancy.
• To utilize multivariate logistic regression analysis incorporating demographic, serologic, and ultrasound variables.
• To identify key predictors of ovarian cancer.

Main Methods:

• Transvaginal ultrasound (B-mode, color, pulsed Doppler) was performed on 171 patients with ovarian tumors.
• Tumors were classified by gray-scale morphology.
• Intratumoral blood flow was assessed using resistance index (RI) and peak systolic velocity (PSV); serum CA 125 levels were measured.

Main Results:

• Unilocular cysts without flow were all benign.
• Among 101 tumors with intratumoral blood flow, multivariate analysis identified serum CA 125 levels (>= 35 U/ml) and PSV (>= 10.4 cm/s) as independent predictors of malignancy.
• A PSV of 10.4 cm/s was the median value in benign tumors.

Conclusions:

• Intratumoral PSV is a powerful tool for distinguishing benign from malignant ovarian tumors.
• Combining PSV with serum CA 125 levels improves the diagnostic accuracy for ovarian malignancy.
• These findings support the integration of Doppler ultrasound parameters in preoperative ovarian tumor assessment.