Background: Prediction of ovarian malignancy by ultrasonographic findings and patient age in the scenario of clinically suspected adnexal masses is a desirable goal.
Methods: Prospective evaluation of clinically suspected adnexal masses was performed with transvaginal ultrasound using real-time, Doppler velocimetry, and color-flow mapping. Continuous ultrasound variables included ovarian volume, the Sassone morphology scale, and Doppler determination of angle-corrected systole, diastole, and time-averaged velocity, in addition to patient age. The Doppler pulsatility index (PI), vessel location, presence of a diastolic notch, and echogenic predominance of the lesion, suggestive of dermoid, also were assessed.
Results: Of 244 women with follow-up, 214 had nonmalignant findings (85 of which were benign neoplasms), and 30 had malignant neoplasms. Age and all ultrasound continuous variables except systole were found to be statistically significant (P < 0.05) between patients with both malignant (N = 30) and nonmalignant masses (N = 214), as well as those with benign (N = 85) and malignant (N = 30) neoplasms. By adding the continuous measures (age [in years], ovarian volume [mL], and Sassone morphology scale [1-15]) and weighting other variables ([-10] x PI, central or septal location [+10], peripheral location [-10], and echogenic [-10]), a receiver operating characteristic curve was generated (area under the curve = 0.91), which was found to be discriminating, predictive, and able to replicate the more complex logistic regression model. Prediction of malignancy was generated from the population-based data of the current study.
Conclusions: The Ovarian Tumor Index, which combines patient age with specific ultrasonographic markers, is an accurate method for predicting ovarian malignancy in the clinical scenario of suspected adnexal masses.
Copyright 1999 American Cancer Society.