Gray-scale ultrasound morphology in the presence or absence of intrauterine fluid and vascularity as assessed by color Doppler for discrimination between benign and malignant endometrium in women with postmenopausal bleeding.

OBJECTIVE To determine if gray-scale ultrasound morphology in the presence or absence of intrauterine fluid and endometrial vascular morphology as assessed by color Doppler ultrasonography can discriminate between benign and malignant endometrium in women with postmenopausal bleeding. METHODS In a prospective study 95 consecutive women with postmenopausal bleeding and endometrial thickness > or = 4.5 mm as measured by transvaginal ultrasound were included. Gray-scale and color Doppler ultrasound examination of the endometrium was performed. The ultrasound examiner characterized the morphology of the endometrium before and during saline infusion and assessed the endometrial vascular tree using a predetermined classification protocol without suggesting a diagnosis. A histopathological diagnosis was obtained by operative hysteroscopy, dilatation and curettage or hysterectomy. RESULTS There were no statistically significant differences in ultrasound findings between benign and malignant endometria of uterine cavities without fluid. Heterogeneous echogenicity, irregular surface, and both heterogeneous echogenicity and irregular surface of a focal lesion (or of the endometrium in the absence of focal lesions) in a uterine cavity filled with fluid (spontaneous or infused) were significantly more common in malignant than in benign endometrium. The sensitivity, false positive rate, positive and negative likelihood ratios of these findings were as follows: heterogeneous echogenicity, 80%, 29%, 2.74, 0.28, P = 0.003; irregular surface, 89%, 33%, 2.70, 0.17, P = 0.002; and both, 78%, 12%, 6.59, 0.25, P < 0.001. Two or more vessels were found in 67% (8/12) of the malignant endometria vs. 51% (40/79) of the benign endometria (non-significant difference). Vascular branching tended to be more common in malignant endometria (10/11; 91%) than in benign endometria (39/61; 64%), P = 0.09. CONCLUSION Heterogeneous echogenicity and an irregular surface of a focal lesion or of the endometrium in a fluid-filled uterine cavity are useful ultrasound criteria for predicting endometrial malignancy. Assessment of vascular morphology using color Doppler ultrasound is of limited--if any--value for discrimination between benign and malignant endometrium.