This document discusses the use of color Doppler sonography and sonohysterography to evaluate abnormal vaginal bleeding. It finds that: 1. Color Doppler sonography (CDS) can detect vessels in both benign and malignant endometrial lesions, but the number of vessels seen correlates with malignancy risk. More vessels were seen in lesions with higher microvessel density. 2. A study of 25 women found CDS could distinguish between endometrial polyps and submucosal fibroids based on vascularity patterns. Polyps typically had a single feeding vessel while fibroids had multiple vessels from the myometrium. Lesions with more vessels >0.5mm on CDS also had higher microvessel

1. 1. Radiological Reasoning: Algorithmic Workup
of Abnormal Vaginal Bleeding with Endovaginal
Sonography and Sonohysterography
diagnostic performance
ofendovaginalsonography for detecting endometrial
pathology in premenopausal women is moderate, with sensitivity
and specificity of 67% and 75%, respectively, using an
endometrial thickness cutoff of > 16 mm [2].
If endovaginalsonography indicates endometrial abnormality,
nonfocal endometrial biopsy should be performed to
exclude cancer or hyperplasia—that is, diffuse endometrial
pathology. Even if cancer is an unlikely possibility, it should
not be missed because it is the one cause of abnormal vaginal
bleeding that is life-threatening. A nonfocal biopsy is a relatively
noninvasive, inexpensive office procedure to evaluate
for endometrial cancer. Once biopsy is negative for cancer or
hyperplasia (a premalignant lesion), workup should continue
to evaluate for the focal benign cause of bleeding.
This patient shows a homogeneous endometrial echo
complex that has a thickness that is within normal limits.
Thus, the vaginal bleeding is unlikely secondary to a diffuse
endometrial process, such as endometrial carcinoma, and a
nonfocal biopsy is not indicated. The next step is to evaluate
for focal causes for bleeding, which are best detected by
sonohysterography.
Sonohysterography
The endometrium measures 1 mm anteriorly and 1 mm
posteriorly. In the endometrial canal is a 2.5 ×1.7 cm homogeneously
hyperechoic lesion, with attachment at the
6-o'clock position on coronal images, and showing a central
stalk on color Doppler imaging
On sonohysterography, endometrial polyps are typically
echogenic, like normal endometrium, and show a narrow attachment
to the normal endometrial lining at its base [9].
Option B is not the best response. A single feeding vessel is
sometimes seen on color Doppler sonography. In contrast,
subendometrial fibroids are typically hypoechoic, like normal
myometrium, and show a broad base of attachment to
themyometrial wall with the normal endometrial lining
overlying its surface [8]. Option C is not the best response.
Fibroids show a hypervascular network of vessels on color
Doppler sonography [10]. Option E is not the best response.
The key to differentiating the two entities is ascertaining the
location of the endometrial lining with regard to the lesion.
The normal endometrial lining underlies the base of a polyp,
whereas it overlies the surface of a fibroid. Option D is not
the best response. Whether a focal lesion is solitary or multiple
onsonohysterography does not distinguish between an
endometrial or subendometrial process

2. 2. Color Doppler Sonography
1.of Endometrial Masses
Color Doppler sonography can depict vessels
with both benign and malignant endometrial
lesions. Therefore, the presence of vessels on
CDS does not seem to directly correlate with a
lesion’s histologic type (benign versus malignant).
However, the number of vessels shown on
CDS correlates with the chance of malignancy.
This parameter did not correlate with microvessel
density. This lack of correlation may be
due to subjective selection of areas to count
microvessels rather than the ability of CDS to
detect smaller vessels.
The correlation of what is depicted on CDS and
vessels seen microscopically is imprecise,
because vessels shown on CDS may not necessarily
be those seen on histologic studies. Larger
vessels (>0.5 mm) may be depicted on CDS if
they have sufficient flow velocities (>1 cm/s),
whereas most vessels seen on histologic studies
are in the area of 10 to 20 m. It can be implied
that masses with greater microvessel density
probably have a greater number of vessels to
supply areas of dense microvascularity, and they
can be depicted on CDS.
Because of the small number of patients in
this study, the importance of our results is limited
and preliminary. However, our results suggest
that CDS can be useful in detection of
endometrial lesions but may not be sufficiently
accurate to differentiate benign from malignant
lesions. We suggest that CDS can be used
as a secondary means to investigate endometria
that appear thickened on TVS. Lesions
containing branching vessels tend to represent
carcinoma, whereas polyps tend to contain
single feeding vessels. Preoperative localization
of the vessels within the pedicle may facilitate
subsequent hysteroscopy, because the
location of the pedicle could be shown beforehand.
A recent study indicates that polyps with
atypia may be distinguished from ones without
atypia by assessment of the flow spectra
within the feeding vessels.6 Because in our
study we did not directly interrogate the
Doppler spectra, we cannot confirm or deny
these findings. In an attempt to optimize
depiction of the number and branches of the
vessels within the endometrial lesion, our
study used power or amplitude Doppler

3. sonography, which precluded simultaneous
spectral analysis.
As the resolution of vascularity improves with
CDS and possible contrast enhancement agents,
CDS may more accurately distinguish endometrial
cancers from polyps. This may be helpful to
determine the proper clinical management,
because polyps are removed hysteroscopically,
whereas cancers require hysterectomy. Color
Doppler sonography currently provides a means
to distinguish thickened endometrial interfaces
due to endometrial polyps or carcinoma from
hematometra, in which flow would not be
shown, thereby distinguishing those patients
who need simple drainage of the luminal contents
from those requiring dilation and curettage
or hysteroscopy. Further investigation with
larger patient populations will help establish the
spectrum of endometrial vascularity seen in
normal postmenopausal women and those taking
medications such as hormone replacement
ortamoxifen.
Color Doppler Sonography of Endometrial Masses
Table 1.Color Doppler Sonographic Results
Endometrial Bilayer No. of Vessels Microvessel Density,
Condition Thickness, mm on CDS No./High-Power Field
Benign polyps (n = 10)* 11–20 (16.2) 1–3 (1.2) 14–116 (79)
Carcinoma (n = 4)* 16–20 (17.4) 2–5 (3.4) 51–109 (76)
Lipoleiomyoma (n = 1) 19 2 116
*Values are range (average).
2. Color Doppler
Sonohysterography
of Endometrial Polyps and
Submucosal Fibroids
1.

4. Of the 25 patients studied, 18 had endometrial
polyps, 3 had intracavitarysubmucosal fibroids,
and 1 had a clot. Three had either secretory or
proliferative endometria. Of the 18 with polyps, 3
had vessels greater than 0.5 mm; all were large
polyps (>10 mm; Figs. 1 and 2). Microvessel
counts ranged from 6 to 102 per HPF and were
greater in the larger polyps (>10 mm). Multiple
vessels were seen on CDS in the 3 patients with a
pedunculatedsubmucosal fibroid, whereas no
flow was seen within the clot (Fig. 3). The
microvessel density in lesions with vessels
greater than 0.5 mm on CDS had a greater average
MVD (60 per HPF) than in those without
visualized vessels (15 per HPF; P = .02).
602 J Ultrasound Med 22:601–604, 2003
Color Doppler Sonohysterography of Polyps and Submucosal Fibroids
Figure
To describe the typical sonographic findings and clinical applications of color Doppler sonohysterography
and to correlate the vascularity of lesions seen on color Doppler sonohysterography to
microvessel density and the presence of vessels greater than 0.5 mm. Methods. Color Doppler sonohysterography
was performed on 25 women with abnormal uterine bleeding. The vascularity (number
of vessels >0.5 mm) and their configuration seen on color Doppler sonohysterography were
compared with those obtained on the excised specimen. Microvessel density and histologic features
were correlated to the visualization of vessels greater than 0.5 mm and their arrangement on color
Doppler sonography.Results.The color Doppler sonographic findings in 18 polyps, 3 submucosal
fibroids, and 1 clot showed distinct vascularity patterns. Polyps typically contained a single feeding vessel,
whereas fibroids had several vessels, which arose from the inner myometrium. Lesions with higher
microvessel density tended to have more vessels greater than 0.5 mm as depicted on color Doppler
sonography. Conclusions.Color Doppler sonohysterography may be useful in distinguishing polyps
fromsubmucosal fibroids based on the vascularity of the lesions. The number of vessels seen on color
Doppler sonography approximates microvessel density within the lesions. Key words: color Doppler
sonography; endometrial polyps; sonohysterography; submucosal fibroids; vascularity
4. Transvaginal Color Doppler
Sonography
Versus Sonohysterography in the
Diagnosis of Endometrial Polyps
Juan Luis Alcázar, MD, María José Galan, MD,
José ÁngelMínguez, MD, Manuel García-Manero, MD
Objective.To compare the diagnostic performance of transvaginal color Doppler sonography (TVCD)
andsonohysterography (SHG) in the diagnosis of endometrial polyps. Methods.Fifty-one women
(mean age, 51 years; range, 27–75 years) with clinical or B-mode sonographic suspicion of endometrial
polyps were included in this prospective study. Transvaginal color Doppler sonography first and
then SHG were performed in all patients. On TVCD, a polyp was suspected when a vascular pedicle
penetrating the endometrium from the myometrium was identified. On SHG, a polyp was suspected
when a focal polypoid lesion was seen within the endometrial cavity. All patients underwent hysteroscopy

5. and endometrial biopsy, the findings of which were used as the criterion standard. Sensitivity
and specificity for TVCD and SHG were calculated and compared by the McNemar test. Results.
Hysteroscopy and endometrial biopsy findings were as follows: endometrial polyps, 41; endometrial
hyperplasia, 3; cystic atrophy, 4; proliferative endometrium, 2; and endometritis, 1. Sensitivity and
specificity for TVCD and SHG were 95% and 80% and 100% and 80%, respectively (McNemar test,
P = .5) Conclusions. Transvaginal color Doppler sonography and SHG had similar performance for
diagnosing endometrial polyps. Key words: color Doppler sonography; endometrial polyp; sonohysterography;
transvaginalsonography.
Received January 14, 2004, from the Department
of Obstetrics and Gynecology, ClínicaUniversitaria
de Navarra, School of Medicine, University of
Navarra, Pamplona, Spain. Revision requested
February 17, 2004. Revised manuscript accepted for
publication February 26, 2004.
Address correspondence and reprint requests to
Juan Luis Alcázar, MD, Department of Obstetrics
and Gynecology, ClínicaUniversitaria de Navarra,
AvenidaPio XII 36, 31008 Pamplona, Spain.
E-mail: