To assess flow characteristics of benign and malignant gynecologic tumors by transvaginal color flow Doppler.Methods:
Records of the Ultrasound Laboratory, Women's Cancer Center, University of Minnesota were analyzed retrospectively. Gray scale findings were recorded as either “diagnostic” or “nondiagnostic.” Color flow assessment was performed on intratumor vessels or ovarian and/or uterine arteries. Flow was recorded as either “absent” or “present.” Spectral analysis allowed determination of the systolic, diastolic, and mean velocities and calculation of the pulsatility and resistance indices. Malignancy was then predicted based upon color flow findings alone, with malignant tumors demonstrating increased color flow and a pulsatility index of at most 1.0 or a resistance index of at most 0.4. Color flow Doppler findings were then recorded as “giving additional useful information” that either confirmed questionable gray scale findings or changed the gray scale sonographic diagnosis, or as “not giving additional information” over the gray scale diagnosis.Results:
Two hundred thirty-one patients had gray scale sonography, and 167 also had color flow Doppler performed. Gray scale sonographic findings were sufficient to make a diagnosis in 156 (93%) of the scans. Color flow Doppler findings added useful information in 49 scans (30%). Increased color flow was highly significant (P<.0001), as was the calculated pulsatility index (P<.02) and resistance index (P<.008), in distinguishing benign from malignant tumors. Ovarian and uterine artery and intratumor assessments of the systolic, diastolic, and mean velocities were not significantly different between the benign and malignant tumors. Regression analysis confirmed the presence or absence of color flow as an independent predictor of malignancy or benignity (P<.0001).Conclusions:
Our large study confirms the overall accuracy of gray scale scanning. When used alone, color flow Doppler—although specific—lacks sensitivity and predictive value as an independent predictor of malignancy. When findings were combined with those obtained from gray scale scanning, sensitivity, specificity, and predictive value were improved to acceptable levels. Significant differences existed between benign and malignant tumors for calculated pulsatility index and resistance index, but neither was sufficiently sensitive, specific, or predictive to be used alone as sole criteria of malignancy prediction. Other flow indices studied (systolic, diastolic, and mean velocities) in general did not differ significantly between groups. Physicians should be cautioned against using color flow findings alone for clinical decision making. We recommend a multiinstitutional study to investigate the multiple vascular assessments to determine the role of color flow Doppler in the preoperative prediction of pelvic tumors and in screening for gynecologic abnormality.