297

ULTRASOUND
CLINICS

Ultrasound Clin 2 (2007) 297–309

Ultrasound of Pelvic Inflammatory

Disease
Mindy M. Horrow, MD, FACR, FAIUMa,b,
Shuchi K. Rodgers, MDa,*, Shabbir Naqvi, MD
a

- Clinical findings - Other imaging: CT

- Ultrasound findings - Other imaging: MRI
Acute pelvic inflammatory disease - Summary
Chronic pelvic inflammatory disease - References
Related findings and differential diagnosis

Pelvic inflammatory disease (PID) is an infection health care system stems more from the major
of the upper genital tract caused by sexually chronic complications than the cost of treating
transmitted disease (STD). In addition to the care the acute infection. Chronic complications include
required during an acute infection, the sequelae of infertility, ectopic pregnancy, and chronic pelvic
infertility, ectopic pregnancy, and chronic pelvic pain. A recent study found the average per-person
pain significantly impact the health care system. lifetime cost of PID ranges from $1060 to $3180
Data from 1990 show an estimated cost for care [4]. Risk factors for PID are related to exposure to
of patients with PID at $4.24 billion annually, STDs and include earlier age at first sexual inter-
with 200,000 hospitalized cases and 1,277,700 course, multiple partners, history of prior STD,
outpatient cases. From 1984 to 1990, hospitaliza- and use of vaginal douche [5]. There is also
tions for PID decreased 25%, with only a slight increased risk from an intra-uterine device (IUD),
rise in outpatient visits [1]. More recent estimates but this is limited to the first few weeks after
from the Centers for Disease Control and Preven- insertion1 [6].
tion approximate 780,000 new cases of acute PID
annually [2]. Although it is unclear if this is a true
Clinical findings
decrease, or just a result of more outpatient care
for patients who have PID, most researchers Most cases of PID are caused by Chlamydia tracho-
estimate that there is a significant cohort with matis or Neisseria gonorrhoeae. Co-infections of these
unrecognized PID. A study in 2004 found the prev- organisms and other bacteria, including: Streptococ-
alence of chlamydial infection in young adults in cus species, Escherichia coli, Haemophilus influenza,
the United States was 4.19%. Women, and in partic- Bacteroides species, Peptostreptococcus and Peptococcus
ular black women, had higher rates of 4.74 and are common. Normally, the major barrier to the
13.95%, respectively [3]. The cost of PID to the assent of both normal vaginal flora and pathogens

a
Department of Radiology, Albert Einstein Medical Center, 5501 Old York Road, Philadelphia,
PA 19141-3098, USA
b
Jefferson Medical College, Thomas Jefferson University, 1025 Walnut Street, Philadelphia, PA 19107, USA
* Corresponding author. Department of Radiology, Albert Einstein Medical Center, 5501 Old York Road,
Philadelphia, PA 19141-3098.
E-mail address: rodgerss@einstein.edu (S.K. Rodgers).
1
The text of this article is adapted from Horrow MM. Ultrasound of pelvic inflammatory disease. Ultrasound
Quarterly 2004;20:171–9; with permission.

1556-858X/07/$ – see front matter ª 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2007.08.008
ultrasound.theclinics.com
298 Horrow et al

is the endocervical canal and its mucus plug. The a screening procedure. Initially, transabdominal
infectious bacteria damage the endocervical canal, sonography, and now transvaginal sonography
permitting organisms to ascend into the uterus. with Doppler have been used in the diagnosis of
Other factors may facilitate the spread of disease. PID.
Cervical ectopy (extension of endocervical colum-
nar epithelium outward beyond the cervix) pro-
Ultrasound findings
duces a larger area that is susceptible to infection.
Cervical ectopy occurs more commonly in teen- Despite the widespread use of sonography in the
agers, who also represent the age group with the diagnosis and management of patients with pelvic
highest incidence of PID. Cervical mucus changes inflammatory disease, there are no large studies
normally during the menstrual cycle, making the evaluating its sensitivity and specificity or overall
cervix more vulnerable to infection at midcycle, usefulness [10–12]. Nonetheless, it is a frequently
when estrogen levels are high and progesterone is ordered study in patients who have unexplained,
relatively low. After ovulation, the mucus becomes acute pelvic pain or in patients with classic symp-
more viscous and less penetrable by both sperm toms of PID in whom an adequate clinical exami-
and bacteria. Bacteria also may gain easier access nation cannot be performed. If demonstration of
to the uterus when the mucus plug is expelled at a pyosalpinx or tubo–ovarian abscess will result in
menstruation [7]. Less frequently, PID occurs as hospitalization, surgery, or follow-up imaging to
a secondary infection from adjacent processes evaluate nonoperative management, then initial
(ie, appendiceal, diverticular, postsurgical ab- sonography is indicated. In addition, many patients
scesses, or puerperal, and postdilatation and curet- are examined initially by generalists who may feel
tage [D & C] complications). Hematogenous spread more confident relying on imaging studies. CT is
is rare, but can occur from tuberculosis. ordered with increasing frequency to rule out alter-
Primary prevention of PID consists of avoiding native diagnoses such as appendicitis and diverticu-
exposure to STDs. Secondary prevention involves litis. Thus, there are more patients in whom CT is
keeping the lower genital infection from ascending the first study to suggest the diagnosis of PID.
into the uterus. This combines disease detection, Though CT is very sensitive for the detection of
treatment, and partner notification. Tertiary pelvic abnormalities, it may not be as specific,
prevention, which involves preventing an upper particularly in differentiating an ovarian from
genital tract infection from leading to tubal dys- a tubal process, and ultrasound often is obtained
function and/or obstruction, has been generally after an abnormal CT study to clarify the patholog-
disappointing [8]. ical process and guide management.
The initial diagnosis of PID, which is based upon As one reviews the imaging literature of the last
a combination of symptoms, pelvic examination, 20 years, it is clear that the sensitivity and specificity
and laboratory studies, is often incorrect. Molander of sonography for PID depends upon the findings
and colleagues [9] confirmed PID by laparoscopy in one considers to be indicators of PID and the qual-
only 67% of women who had clinically suspected ity of the equipment and the sonographer. Transab-
PID. Laparoscopy, however, is not suitable for dominal imaging is suitable for determining overall

Fig. 1. (A) Longitudinal, transabdominal image of the uterus at the time of presentation demonstrates an ante-
verted, anteflexed uterus with an indistinct endometrial stripe and indistinct margins. (B) Re-examination 3 days
later after treatment shows a distinct endometrial stripe and uterine margins. (From Horrow MM. Ultrasound of
pelvic inflammatory disease. Ultrasound Quarterly 2004;20:171–9; with permission.)
 Ultrasound of Pelvic Inflammatory Disease 299

Fig. 4. Transverse, transvaginal view demonstrating

Fig. 2. Sagittal, transvaginal view of the uterus dem- complex fluid with low level echoes within the cul
onstrates prominent echogenic fat (arrows) adjacent de sac from a ruptured hemorrhagic ovarian cyst.
to the anterior, superior aspect of the uterine fundus. The fimbriated portion of a normal fallopian tube
(T) is visualized also.

extent of disease. Transvaginal sonography with

color or power Doppler, however, has significantly sonographer to correlate the patient’s symptoms
improved the ability to detect subtle abnormalities with the imaging findings, by eliciting focal tender-
such as slightly swollen fallopian tubes and to ness upon examination of the affected fallopian
distinguish whether the ovary is involved in an tubes and ovaries. This is particularly valuable
abscess [13]. In addition, many patients who have when the imaging findings are subtle.
acute PID cannot tolerate an adequately filled
urinary bladder; therefore transabdominal sonogra-
Acute pelvic inflammatory disease
phy is suboptimal [14]. Papers describing the find-
ings of PID in the pretransvaginal era are therefore In a patient without an abscess or pyosalpinx, the
of limited use. manifestations of PID are often subtle and some-
In medical imaging, the ability to make the times nonspecific. Mild enlargement or indistinct-
correct diagnosis depends upon knowledge of the ness of the uterus can be very difficult to appreciate
possible manifestations of the disease and correla- and often only confirmed if a follow-up examina-
tion with the clinical findings. Transvaginal imag- tion is performed (Fig. 1). If the patient has some
ing is suited uniquely in both these regards for fat in the pelvis, inflammatory changes will be man-
diagnosing PID. The higher frequency transvaginal ifest as increased echogenicity and prominence of
probes allow one to observe the more subtle this fat (Fig. 2). It is the authors’ experience that sub-
changes of salpingitis and oophoritis. Furthermore, tle changes in uterine size and of the adjacent fat are
the act of transvaginal imaging itself permits the

Fig. 5. Transverse, transvaginal view of the right

Fig. 3. Sagittal, transvaginal view demonstrates fluid adnexa in a patient with pelvic inflammatory disease
in the endometrial canal consistent with but not demonstrates an enlarged right ovary (volume 23 cc)
specific for endometritis in this patient with pelvic with multiple tiny peripheral cysts. There is also an
inflammatory disease. adjacent thickened fallopian tube (arrows).
300 Horrow et al

Fig. 8. Transverse, transvaginal view of a fallopian

tube (arrows) with a thick, echogenic wall, contain-
Fig. 6. Transverse, transvaginal image demonstrates ing complex fluid, representing a pyosalpinx.
a normal fallopian tube outlined by fluid from a rup-
tured hemorrhagic cyst.
and increasing the number and size of cysts. The
margins of the ovaries also may become indistinct
appreciated better transabdominally because of the in PID.
wider field of view. Both fluid in the endometrium Sonographic findings of the fallopian tubes are
and fluid in the cul de sac, common findings in the most striking and specific landmarks of PID.
PID, can be found in normal patients and those Normal fallopian tubes are difficult to visualize
with alternative diagnoses (Figs. 3 and 4). In addi- on routine transvaginal sonography, unless
tion, some patients who have PID may have little outlined by fluid (Fig. 6). The normal fallopian
or no free fluid [11]. Nonetheless, the presence of tube is approximately 10 cm long and runs along
even small amounts of complex pelvic fluid is worri- the superior margin of the broad ligament. It can
some for free pus. Careful attention to gain control be divided into four segments from the proximal
settings is necessary to visualize internal echoes. uterine end to the distal fimbriated end as follows:
Two groups have reported the finding of enlarged interstitial, isthmic, ampullary, and infundibular.
ovaries with increased numbers of small cysts, a so- The fimbriated portion of the infundibular segment
called polycystic ovary appearance, to correlate with attaches to the ovary and is open to the peritoneal
PID (Fig. 5). Cacciatore and colleagues [12] and cavity [16]. The diameter of the tube typically varies
Golden and colleagues [15] both found larger- between 1 and 4 mm, and in the authors’ experi-
than-normal ovarian volumes in their patients ence is thickest in the infundibular segment.
who had PID proven by endometrial biopsy or When inflamed, the fallopian tube swells, and
laparoscopy. These authors have found increased the walls and endosalpingeal folds thicken,
ovarian volume to be a useful sign of PID, having allowing visualization with ultrasound (Fig. 7).
documented decreasing volumes on follow-up This initial stage corresponds to salpingitis, a thick-
examinations. A reasonable explanation for this ened fallopian tube without intraluminal pus. On
reactive polycystic change is that the inflammation ultrasound, it can appear as an indistinct, elon-
increases ovarian volume by thickening the stroma gated, noncystic mass in close proximity to the

Fig. 7. Transverse view of the pelvis with thickened

left fallopian tube (arrow), not filled with fluid.
(From Horrow MM. Ultrasound of Pelvic Inflamma- Fig. 9. Transverse, transvaginal view of a slightly
tory Disease. Ultrasound Quarterly 2004;20:171–9; dilated, fluid-filled fallopian tube with thickened endo-
with permission.) salpingeal folds demonstrating the cog wheel sign.
 Ultrasound of Pelvic Inflammatory Disease 301

Fig. 10. (A) Cross section of significantly dilated tube (arrows) containing echogenic fluid that outlines thickened
endosalpingeal folds, demonstrating the cog wheel sign. (B) Photomicrograph of same tube in cross section with
thickened folds projecting into a dilated lumen. (From Horrow MM. Ultrasound of pelvic inflammatory disease.
Ultrasound Quarterly 2004;20:171–9; with permission.)

ovary, but separated from it. The endosalpingeal not affected initially, the disease may spread there
folds are difficult to discern given the lack of intra- secondarily. When both tubes are inflamed and
luminal fluid as a contrast. The pus may flow freely occluded, the entire complex typically takes on
from the tube into the peritoneum, preventing dis- a U shape as it fills the cul de sac, extending from
tention and detection of a thickened tubal wall. one adnexal region to the other. The lateral and
As the lumen occludes distally, the fallopian tube posterior borders of the uterus become obscured,
distends and fills with complex fluid, resulting in and individual tubes and ovaries cannot be
a pyosalpinx. Various appearances result, as distinguished.
described by Timor-Tritsh and colleagues [17]. Color and power Doppler may show increased
The tube becomes ovoid or pear-shaped, filling flow (hyperemia) in the walls and incomplete septi
with fluid that may be anechoic or echogenic, of the inflamed tubes (Fig. 16). In the acute phase
with layers. The wall becomes thickened, greater of the infection, the mean resistive and pulsatility
than or equal to 5 mm (Fig. 8), and incomplete indices may be low, measuring 0.5 (standard devia-
septi are common as the tube folds back upon itself. tion [SD] 5 0.05) and 0.79 (SD 5 0.12). With treat-
If the distended tube is viewed in cross section, it ment, these values increased to 0.63 and 1.17,
may demonstrate the cog wheel sign (Figs. 9 and respectively [18]. Some authors have shown that
10A), because of the thickened endosalpingeal folds
(Fig. 10B) [17]. Typically the swollen fallopian
tubes extend posteriorly into the cul de sac, rather
than extending superiorly and anterior to the uterus
as large ovarian tumors tend to do. Fluid debris
levels often are visualized in the dilated tubes
(Fig. 11), and very rarely gas/fluid levels or bubbles
of gas (Fig. 12).
As the disease progresses, the ovary can become
involved. Theoretically, a defect in the ovary at the
time of ovulation allows bacteria to enter, spreading
the infection. When the ovary adheres to the tube,
but remains visualized, this indicates a tubo–
ovarian complex (Fig. 13). A tubo–ovarian abscess
is the result of a complete breakdown of ovarian
and tubal architecture such that separate structures
no longer are identified (Fig. 14). Without treat-
ment, a tubo–ovarian abscess can rupture, resulting
in peritonitis and multiple intra-abdominal Fig. 11. Oblique, transvaginal view of a dilated right
abscesses (Fig. 15). If the contralateral side was fallopian tube containing a fluid–pus level.
302 Horrow et al

Fig. 12. (A) Dilated left fallopian tube filled with fluid and multiple bright echogenic foci representing bubbles
of gas. (B) CT image through the pelvis in the same patient shows dilated tubes (arrows) filling the cul de sac.
The left tube contains a gas/fluid level. (From Horrow MM. Ultrasound of pelvic inflammatory disease. Ultra-
sound Quarterly 2004;20:171–9; with permission.)

patients who responded to conservative, medical following hysterectomy if the fallopian tubes are
treatment tended to have a higher resistive index left in to protect the vascular supply to the ovary,
(.6  .15) than those who required surgery (.52  and primary or secondary tumors of the fallopian
.08), although there was significant overlap be- tubes [16]. Several specific ultrasound findings
tween the two groups [19]. It is the authors’ experi- can help distinguish a hydrosalpinx from other cys-
ence that color Doppler imaging can be useful to tic adnexal lesions. A hydrosalpinx tends to be an-
differentiate PID from tumors or masses, but spe- echoic, more tubular, and often demonstrates the
cific resistive and pulsatility indices are not helpful. incomplete septa sign (Fig. 17). The tubal wall is
Decreased flow on follow-up imaging may be useful thin, less than 5 mm, and in cross section demon-
to assess for response to therapy. strates the beads-on-a-string sign (Fig. 18) [17].
These beads are 2 to 3 mm hyperechoic nodules
projecting from the wall, representing remnants of
Chronic pelvic inflammatory disease the endosalpingeal folds. If color flow is detected
Chronic PID typically results in a hydrosalpinx in a hydrosalpinx, it tends to be less exuberant
from accumulation of fluid caused by occlusion of than in acute PID. Molander and colleagues [9]
the tube distally or at both ends. Other causes of hy- found a higher pulsatility index in patients who
drosalpinx, however, include tubal ligation, had a chronic hydrosalpinx (1.5  .1) than with

Fig. 13. Tubo–ovarian complex. Sagittal, transvaginal Fig. 14. Transverse, transvaginal image of the pelvis
image of the left adnexa demonstrates an ovary (O) demonstrates a right tubo–ovarian abscess (TOA)
with ill-defined borders, surrounded by a thickened and a left tubo–ovarian complex consisting of a pyo-
fallopian tube (T) containing fluid (F) and pus (P). salpinx (T) and the adjacent left ovary (OV).
 Ultrasound of Pelvic Inflammatory Disease 303

Fig. 17. Sagittal, transvaginal image of a hydrosalpinx,

demonstrating an incomplete septum (arrow).

when fluid from a ruptured ovarian cyst is trapped

around the ovary by adhesions (Fig. 19).
Fig. 15. Coronal enhanced CT scan in a different Sonographic changes can occur fairly rapidly af-
patient demonstrating peritoneal inflammation ter treatment of PID. Minor findings such as com-
(white arrows) and multiple abscesses (black arrows)
plex fluid and inflammation in the surrounding
from a ruptured tubo–ovarian abscess.
fat can resolve in a few days. A pyosalpinx can
change to a hydrosalpinx and possibly resolve
acute PID (.84  .04). Occasional patients with over a few weeks to several months. Interestingly,
prior PID may demonstrate a peritoneal inclusion Taipale and colleagues [21] found that 9 of 55 pa-
cyst. This diagnosis is made when the ovary is sur- tients who had clinical PID and an initial, normal
rounded by a loculated fluid collection with thin sonogram developed a hydrosalpinx with time. If
septations [20]. A peritoneal inclusion cyst forms a pyosalpinx does not resolve or develops into

Fig. 16. Color and pulsed Doppler image of a left pyosalpinx demonstrates low resistance arterial flow (RI 5 .49).
304 Horrow et al

Fig. 18. (A) Hydrosalpinx. Lengthwise, transvaginal view of a dilated, fluid-filled fallopian tube demonstrating
residual endosalpingeal folds (arrows). (B) Cross section of the same tube shows the beads on a string sign.

a hydrosalpinx, it probably signifies an incom- on ultrasound in nine patients. Scattered cases of

pletely treated infection. Fitz-Hugh-Curtis syndrome on CT have demon-
strated increased enhancement of the peritoneal
Related findings and differential diagnosis surface of the anterior aspect of the liver, gallblad-
Patients who have severe PID can develop a reactive der wall thickening, and a transient hepatic attenu-
ileus (Fig. 20), which can be visualized during ation difference [24,25].
sonographic evaluation of the pelvis. Perihepatitis Though the combination of sonographic and
associated with PID is known as Fitz-Hugh-Curtis clinical findings is often quite specific for PID, there
Syndrome, which occurs in 3% to 10% of patients are several other common diagnoses in the differen-
who have PID [22]. The acute right upper quadrant tial diagnosis. The most common alternative diag-
pain often overshadows the pelvic pain because of noses with findings which simulate PID by the
acute salpingitis. Inflammatory exudates in the presence of an indistinct uterus and complex pelvic
pouch of Douglas spread along the peritoneal sur- fluid, are ruptured hemorrhagic cyst and endome-
face to the anterior surface of the liver by means trioma. Perforated appendicitis and ruptured
of the paracolic gutter (Fig. 21). Often alternative tubo-ovarian abscess may present with indistin-
diagnoses such as cholecystitis are considered first, guishable findings of peritonitis and intra-abdomi-
and imaging of the right upper quadrant may be nal abscesses. Other tubular structures in the pelvis
requested. Shoenfeld and colleagues [23] found which bear a resemblance to dilated fallopian tubes
thickening of the right anterior extrarenal fascia but should be distinguishable from them include

Fig. 19. Sagittal image of the right adnexa demon-

strating loculated fluid with thin septations, sur- Fig. 20. Transabdominal image of the right pelvis in
rounding the right ovary compatible with a patient with pelvic inflammatory disease demon-
a peritoneal inclusion cyst. The patient has a history strates dilated loops of small bowel without peristal-
of pelvic inflammatory disease. sis, compatible with a reactive ileus.
 Ultrasound of Pelvic Inflammatory Disease 305

A hydrosalpinx can mimic a cystic ovarian mass.

This occurs when a significantly scarred hydrosal-
pinx has multiple septa without connecting
compartments. Conversely, cystic ovarian tumors
seldom have incomplete septa. Large para-tubal or
para-ovarian cysts can also appear similar to a hydro-
salpinx. For example, a hydatid of Morgagni is a be-
nign cystic structure arising from a müllerian duct
vestige adjacent to the fimbrial end of the fallopian
tube (Fig. 25). Identification of the ovary separate
from the hydrosalpinx allows differentiation from
an ovarian mass. MRI or CT may aid in difficult cases.
Torsion of the fallopian tube is a rare condition
that can occur with PID or other pre-disposing
conditions. These include hydrosalpinx, tubal
tortuosity, tubal ligation, tumor, or extrinsic causes
such as a paratubal mass, adjacent adhesions, or
Fig. 21. Fitz-Hugh-Curtis syndrome in a patient with enlargement of the uterus compressing the tube.
pelvic inflammatory disease. Coronal unenhanced The clinical presentation is similar to ovarian tor-
CT image demonstrates peritoneal inflammation sion. Sonography typically shows a thick-walled dis-
(white arrows) extending to the inferior liver capsule tended fallopian tube, which may contain
(black arrow). The patient presented with right upper hemorrhage. Differentiation of acute PID with tubal
quadrant pain. pus from tubal torsion with hematosalpinx may be
difficult. Since tubal blood supply is from both ovar-
ian and uterine vessels, the presence of flow on color
an inflamed appendix (Fig. 22), a hydroureter,
Doppler does not exclude the diagnosis. Ultimately,
prominent pelvic veins (Fig. 23), and varices.
the prospective diagnosis of tubal torsion is very dif-
Septic ovarian vein or pelvic vein thrombophle-
ficult and is most often made pathologically [16,27].
bitis may present with a similar clinical picture as
PID. However, it most commonly occurs postpar-
tum, but can be associated with pelvic trauma,
Other imaging: CT
PID, or recent pelvic surgery. In the majority of
cases, it occurs on the right side, and during the first When significant sonographic findings are detected
week postpartum. On sonography, a tubular in the pelvis, sonography should be extended into
anechoic or hypoechoic structure is identified, the abdomen. Complex pelvic fluid may ascend
extending superiorly from the adnexa, with absence into the flanks and Morison’s pouch. Hydroureter-
of flow on Doppler imaging (Fig. 24) [26]. onephrosis is not uncommon. Patients with nausea

Fig. 22. (A) Transabdominal view of the right adnexa demonstrates a tubular, blind ending structure with low-
level internal echoes. Patient had surgically proven appendicitis. (B) Transverse view of the same dilated appen-
dix with surrounding echogenic fat consistent with inflammation.
306 Horrow et al

Fig. 23. (A) Sagittal, transvaginal view of the right adnexa demonstrates a tubular structure adjacent to the right
ovary (O). (B) Corresponding color Doppler image demonstrates that the tubular structure represents a promi-
nent iliac vein.

and vomiting often demonstrate dilated loops of CT is also beneficial in detecting involvement of ad-
small bowel with minimal peristalsis, indicative of jacent structures such as: small and/or large bowel
an ileus. ileus or obstruction, ureteral obstruction, secondary
While ultrasound remains the imaging modality inflammation of the appendix and inflammation of
of choice in cases of suspected PID, if the symptoms the greater omentum. The extent of a ruptured
are more generalized and non-specific, CT is often tubo-ovarian abscess is better appreciated with CT.
ordered first. CT should be performed with both In contrast to ultrasound however, it is more diffi-
oral and intravenous contrast. While many of the cult to differentiate a pyosalpinx from a tubo-ovar-
classic sonographic findings including enlarged ian complex or abscess by CT. Similarly mildly
ovaries, dilated tubes and free-fluid are equally dilated tubes may go unrecognized on CT. A sub-
well seen on CT, the mild inflammatory changes group of patients with an intra-uterine contracep-
of PID may be better appreciated on CT than ultra- tive device in place may develop a particular type
sound. Mild pelvic edema causes thickening of the of subacute or indolent form of PID. They are prone
utero–sacral ligaments and haziness of pelvic fat to infection with Actinomyces israelii, leading to
[28]. Periovarian stranding and enhancement of a more chronic, suppurative infection which may
the peritoneum, endometrium and endocervical ca- simulate a neoplasm with carcinomatosis on CT
nal are well visualized with CT (Figs. 26 and 27). [29].

Fig. 24. (A) Axial enhanced CT of the pelvis in 13-year-old recently postpartum patient shows bilateral tubular
low density structures with enhancing walls, compatible with thrombosed, septic pelvic veins. (B) Corresponding
transverse color Doppler image shows hypoechoic thrombus in a left pelvic vein.
 Ultrasound of Pelvic Inflammatory Disease 307

Fig. 27. CT scan through superior aspect of pelvis

Fig. 25. Transverse, transvaginal view of the left shows distended bladder (B) and inflammation of
adnexa demonstrates a paratubal cyst (Hydatid of peritoneal fat (arrow). The anterior peritoneum is
Morgagni) arising from the fimbriated portion of thickened and enhancing, and small collections of
the left fallopian tube. fluid are present (F). (From Horrow MM. Ultrasound
of pelvic inflammatory disease. Ultrasound Quarterly
2004;20:171–9; with permission.)

Other imaging: MRI

ultrasound cases (Fig. 28). For example, MRI can
Scattered reports of MR imaging for PID show find-
distinguish hematosalpinx from pyosalpinx, both
ings similar to CT. Pelvic inflammation appears as
of which can look similar on ultrasound.
ill-defined hyperintense areas on fat-suppressed
In a study comparing ultrasound and MRI in
T2 weighted images and marked enhancement on
a group of 21 patients with PID proven at laparos-
gadolinium-enhanced fat-suppressed T1-weighted
copy, sensitivity and specificity for PID was 95%
images. Contents of a pyosalpinx or abscess are
and 89% by MRI and 81% and 78% for ultra-
slightly hyperintense on T1-weighted images and
sound [30]. The authors felt that MRI was better
slightly hypointense on T2-weighted images relative
than ultrasound for distinguishing other causes
to urine due to hemorrhage or debris. The superior
of cystic adnexal masses when a dilated tube was
tissue contrast and multiplanar capability of MRI is
not present. Also, MRI was more sensitive for
occasionally useful in problem solving difficult
small amounts of fluid throughout the pelvis.
Nonetheless, the significantly lower cost and eas-
ier availability of ultrasound makes it the initial
imaging study of choice, with MRI used selectively
for problem solving in complex or equivocal
cases.

Summary
Though the true sensitivity and specificity of ultra-
sound for PID are unknown, this study is frequently
ordered. Awareness of the subtle findings of PID,
particularly those that distinguish a dilated
fallopian tube from other cystic adnexal masses
(‘‘incomplete septa,’’ ‘‘cog wheel,’’ and ‘‘beads-on-
a-string’’ signs) will allow the interpreter to be
Fig. 26. CT scan through pelvis shows thickening of
more accurate. Transvaginal scanning allows one
the broad ligaments (thin arrows) and thickening
and enhancement of the peritoneum in the cul de to correlate imaging findings with symptoms. As
sac (thick arrows). The cul de sac contains fluid and CT is used with increasing frequency, imagers
dilated fallopian tubes. (From Horrow MM. Ultra- must be able to appreciate the findings of PID on
sound of pelvic inflammatory disease. Ultrasound this modality and when to correlate with
Quarterly 2004;20:171–9; with permission.) sonography.
308 Horrow et al

Fig. 28. (A) Transverse, transvaginal image demonstrating a cystic right adnexal mass with fine, uniform low-
level echoes and peripheral echogenic mural nodule, highly suggestive of an endometrioma. (B) Coronal T2-
weighted MR image shows lumen of the same cystic mass continuous with the adjacent fallopian tube (arrow),
indicating that it is tubal in origin. Surgical pathology proved this structure to be a pyosalpinx.

[12] Cacciatore B, Leminen A, Ingman-Friberg S, et al.

References
Transvaginal sonographic findings in ambula-
[1] Washington AE, Katz P. Cost of and payment tory patients with suspected pelvic inflammatory
source for pelvic inflammatory disease: trends disease. Obstet Gynecol 1992;80:912–6.
and projections, 1983 through 2000. JAMA [13] Bulas DI, Ahlstrom PA, Sivit CJ, et al. Pelvic
1991;266:2565–9. inflammatory disease in the adolescent:
[2] Rein DB, Kassler WJ, Irwin KL, et al. Direct med- comparison of transabdominal and transvagi-
ical cost of pelvic inflammatory disease and its nal sonographic evaluation. Radiology 1992;
sequelae: decreasing but still substantial. Obstet 183:435–9.
Gynecol 2000;95:397–402. [14] Bellah RD, Rosenberg HK. Transvaginal ultra-
[3] Miller WC, Ford CA, Morris M, et al. Prevalence sound in a children’s hospital: is it worthwhile?
of chlamydial and gonococcal infections among Pediatr Radiol 1991;21:570–4.
young adults in the United States. JAMA 2004; [15] Golden N, Cohen H, Gennari G, et al. The use of
291:2229–36. pelvic ultrasonography in the evaluation of
[4] Yeh JM, Hook EW, Goldie SJ. A refined estimate adolescents with pelvic inflammatory disease.
of the average lifetime cost of pelvic inflamma- Am J Dis Child 1987;141:1235–8.
tory disease. Sex Transm Dis 2003;30:369–78. [16] Benjaminov O, Atri M. Sonography of the abnor-
[5] Aral SO, Mosher WD, Cates W Jr. Self-reported mal fallopian tube. AJR Am J Roentgenol 2004;
pelvic inflammatory disease in the United States, 183:737–42.
1988. JAMA 1991;266:2570–3. [17] Timor-Tritsch IE, Lerner JP, Monteagudo A, et al.
[6] Barrett S, Taylor C. A review on pelvic inflamma- Transvaginal sonographic markers of tubal
tory disease. Int J STD AIDS 2005;16:715–21. inflammatory disease. Ultrasound Obstet Gyne-
[7] Rice PA, Schachter J. Pathogenesis of pelvic col 1998;12:56–66.
inflammatory disease: what are the questions? [18] Tinkanen H, Kujansuu E. Doppler ultrasound
JAMA 1991;266:2587–93. findings in tubo–ovarian infectious complex.
[8] Washington AE, Cates W, Wasserheit JN. Prevent- J Clin Ultrasound 1993;21:175–7.
ing pelvic inflammatory disease. JAMA 1991;266: [19] Tepper R, Aviram R, Cohen N, et al. Doppler
2574–80. flow characteristics in patients with pelvic
[9] Molander P, Sjoberg J, Paavonen J, et al. Transva- inflammatory disease: responders versus nonre-
ginal power Doppler findings in laparoscopically sponders to therapy. J Clin Ultrasound 1998;
proven acute pelvic inflammatory disease. Ultra- 26:247–9.
sound Obstet Gynecol 2002;17:233–8. [20] Horrow MM, Brown KJ. Femscan: multiloculated
[10] Patten RM, Vincent LM, Wolner-Hanssen P, et al. pelvic cyst. Journal of Women’s Imaging 2002;4:
Pelvic inflammatory disease: endovaginal sonog- 89–90.
raphy with laparoscopic correlation. J Ultra- [21] Taipale P, Tarjanne H, Ylostalo P. Transvaginal so-
sound Med 1990;9:681–9. nography in suspected pelvic inflammatory dis-
[11] Boardman LA, Peipert JF, Brody JM, et al. Endovagi- ease. Ultrasound Obstet Gynecol 1995;6:430–4.
nal sonography for the diagnosis of upper genital [22] Ignacio E, Hill M. Ultrasound of the acute female
tract infection. Obstet and Gynecol 1997;90:54–7. pelvis. Ultrasound Q 2003;19:86–98.
 Ultrasound of Pelvic Inflammatory Disease 309

[23] Schoenfeld A, Fisch B, Cohen M, et al. Ultra- by CT and Sonography. AJR Am J Roentgenol
sound findings in perihepatitis associated with 1998;171:487–90.
pelvic inflammatory disease. J Clin Ultrasound [27] Gross M, Blumstein SL, Chow LC. Isolated fallo-
1992;20:339–42. pian tube torsion: a rare twist on a common
[24] Tsubuku M, Hayashi S, Terahara A, et al. Fitz- theme. AJR Am J Roentgenol 2005;185:1590–2.
Hugh-Curtis syndrome: linear contrast enhance- [28] Sam JW, Jacobs JE, Birnbaum BA. Spectrum of
ment of the surface of the liver on CT. J Comput CT findings in acute pyogenic pelvic inflamma-
Assist Tomogr 2002;26:456–8. tory disease. RadioGraphics 2002;22:1327–34.
[25] Pickhardt PJ, Fleishman MJ, Fisher AJ. Fitz- [29] Lee I-J, Ha HK, Park CM, et al. Abdominopelvic
Hugh-Curtis syndrome: multidetector CT find- actinomycosis involving the gastrointestinal
ings of transient hepatic attenuation difference tract: CT features. Radiology 2001;220:76–80.
and gallbladder wall thickening. AJR Am J Roent- [30] Tukeva TA, Aronen HJ, Karjalainen PT, et al. MR
genol 2003;180:1605–6. imaging in pelvic inflammatory disease: compar-
[26] Quane LK, Kidney DD, Cohen AJ. Unusual ison with laparoscopy and US. Radiology 1999;
causes of ovarian vein thrombosis as revealed 210:209–16.