Müllerian Duct Anomalies: Imaging and Clinical Issues

Radiology
State of the Art
Robert N. Troiano, MD
Shirley M. McCarthy, MD,
PhD
Index terms:
Genitourinary system, abnormalities
Genitourinary system, MR,
85.121411
Genitourinary system, radiography,
85.1282
Genitourinary system, US, 85.12981,
85.12983
State of the Art
Uterus, abnormalities, 854.1411,
854.1413, 854.14783, 854.14784
Published online before print
10.1148/radiol.2331020777
Radiology 2004; 233:19 –34
Abbreviations:
DES ⫽ diethylstilbestrol
HSG ⫽ hysterosalpingography
1
From the Departments of Radiology
and Obstetrics and Gynecology, Weill
Medical College of Cornell University,
1300 York Ave, New York, NY 10021
(R.M.T.), and Department of Radiology, Yale University School of Medicine,
New Haven, Conn (S.M.M.). Received
June 25, 2002; revision requested August 22; final revision received August
25, 2003; accepted September 29.
Address correspondence to R.N.T.
(e-mail: [email protected]).
Authors stated no financial relationship to disclose.
©
RSNA, 2004
Müllerian Duct Anomalies:
Imaging and Clinical Issues1
While estimates of the frequency of müllerian duct anomalies vary widely owing to
different patient populations, nonstandardized classification systems, and differences in diagnostic data acquisition, these anomalies are clinically important, particularly in women who present with infertility. An understanding of the differences
between these uterovaginal anomalies, as outlined in the most widely accepted
classification system—that published by the American Fertility Society (AFS) in
1988 —is imperative given the respective clinical manifestations, different treatment
regimens, and prognosis for fetal salvage. Although the AFS classification system
serves as a framework for description of anomalies, communication among physicians, and comparison of therapeutic modalities, there often is confusion about
appropriate reporting of certain anomalies, particularly those with features of more
than one class. Many of the anomalies are initially diagnosed at hysterosalpingography and ultrasonography; however, further imaging is often required for definitive
diagnosis and elaboration of secondary findings. At this time, magnetic resonance
imaging is the study of choice because of its high accuracy and detailed elaboration
of uterovaginal anatomy. Laparoscopy and hysteroscopy are reserved for women in
whom interventional therapy is likely to be undertaken.
©
RSNA, 2004
The true incidence and prevalence of müllerian duct anomalies are difficult to assess.
Examination of different patient populations, nonstandardized classification systems, and
differences in diagnostic data acquisition have resulted in widely disparate estimates, with
a reported prevalence that ranges from 0.16% to 10% (1–9). A reflection of selection bias,
a prevalence of 0.4% has been reported in women being evaluated with ultrasonography
(US) because of nonobstetric indications (2), while a prevalence of 8%–10% has been
reported in women being evaluated with hysterosalpingography because of recurrent
pregnancy loss (6,7). The overall data suggest that the prevalence both in women with
normal fertility (1,10,11) and in women with infertility (10 –13) approximates 1%, and the
prevalence in women with repeated pregnancy loss approximates 3% (10,11,14 –16).
While the majority of women with müllerian duct anomalies have little problem conceiving, they have higher associated rates of spontaneous abortion, premature delivery, and
abnormal fetal lie and dystocia at delivery (17,18). Most studies report an approximate
frequency of 25% for associated reproductive problems, compared with 10% in the general
population (18 –20). Primary infertility in these women usually has an extrauterine cause and
is not generally attributable to müllerian duct anomalies alone (18,21). In addition, cervical
incompetence has been reported to be associated with these anomalies (3,14,22).
The majority of müllerian duct anomalies are considered to be sporadic or multifactorial
in nature; however, polygenic and genetic patterns of inheritance have been described in
the expression of these anomalies (23,24). Extrauterine and intrauterine environmental
factors, such as exposure to ionizing radiation, intrauterine infections, and drugs with
teratogenic effects such as thalidomide and diethylstilbestrol (DES), can also cause defects
of the developing fetal genital tracts (18).
EMBRYOLOLOGY
At 6 weeks of development, the male and female genital systems are indistinguishable in
appearance, constituting two sets of paired ducts: the paramesonephric (müllerian) ducts
and the mesonephric (wolffian) ducts. In the absence of the testis-determining factor of
19
Radiology
ESSENTIALS
●
Congenital uterine anomalies are clinically important, especially in women
who present with infertility.
●
Precise imaging characterization is essential, given the different clinical manifestations, treatment regimens, and
prognoses for fetal salvage.
●
MR imaging is currently the modality
of choice for further evaluation of uterine anomalies.
the Y chromosome, the mesonephric
ducts begin to degenerate and form a matrix for the developing paramesonephric
ducts. Synchronously, the paramesonephric ducts develop bidirectionally
along the lateral aspects of the gonads.
The proximal segments of the uterovaginal canal, derived from coelomic epithelium, remain unfused and open into the
peritoneal cavity to form the fallopian
tubes. The distal segments, induced by or
derived from the adjacent mesonephric
ducts, progress caudomedially and join
each other before contacting the posterior aspect of the pelvic urethra at the
level of the sinusal tubercle. These distal
segments of the uterovaginal canal give
rise to the uterus and upper four-fifths of
the vagina.
Initially separated by a septum, at 9
weeks the paramesonephric ducts fuse at
their inferior margin forming the single lumen of the uterovaginal canal. Regression
of the uterine septum has been proposed to
be a result of apoptosis, mediated by the
Bcl2 gene (25). Absence of this gene has
been implicated in persistence of the septum. The classic theory of unidirectional
regression hypothesizes that the septum
regresses from the caudal to cranial aspect
of the uterovaginal canal, with the uterus
initially bicornuate in configuration. However, an alternative bidirectional theory
has been proposed in which it is hypothesized that the process proceeds simultaneously in both the cranial and the caudal
directions (26). This would explain anomalies such as a complete septum with a
duplicated cervix or isolated vertical upper
vaginal septum in an otherwise unremarkable uterus.
At week 12, the uterus exhibits its normally developed configuration: a fused
external uterine contour of the myometrium and a triangular-shaped endome20
䡠
Radiology
䡠
October 2004
trium. Because the fallopian tubes are derived from a different cellular origin than
are the uterus and mid- to upper vagina,
they are rarely involved in müllerian
duct anomalies.
During formation of the uterovaginal
canal, the sinusal tubercle thickens and
forms the sinovaginal bulbs of the primitive urogenital sinus, which gives rise to
the lower 20% of the vagina. The uterovaginal canal remains separated from the
sinovaginal bulbs by the horizontal vaginal plate. The vaginal plate elongates
during the 3rd–5th month, and its interface with the urogenital sinus forms the
hymen, which usually ruptures during
the perinatal period (27–29).
The urinary and genital systems both
arise from a common ridge of mesoderm
arising along the dorsal body wall, and
both rely on normal development of the
mesonephric system. The ureters, renal
calices, and collecting tubules are formed
from the ureteral bud, which arises from
the mesonephric ducts, which also induce formation of the kidneys. Hence,
abnormal differentiation of the mesonephric and paramesonephric ducts may
also be associated with anomalies of the
kidneys. Renal agenesis is the most common associated anomaly, although crossed
renal ectopy, cystic renal dysplasia, and duplicated collecting systems have all been
described (26 –30).
The ovaries arise from the mesenchyme
and epithelium of the gonadal ridge and
are not influenced by the formation of the
mesonephric or paramesonephric ducts.
The undifferentiated gonads are induced
to develop by primordial germ cells that
migrate from the yolk sac to the dorsal
mesenchyme at 5 weeks. These germ cells
induce cells of the mesonephros to form
genital ridges, which in turn form primitive sex cords. If germ cells do not develop
in the region of the gonads, the gonads do
not form. Hence, ovarian development is a
separate process from the formation of the
uterovaginal canal and is not usually associated with müllerian duct anomalies (27).
IMAGING TECHNIQUES
Hysterosalpingography (HSG) is indicated in the early stages of evaluation of
the infertile couple (31,32). The examination provides a morphologic assessment of the endometrial and endocervical canals and supplies important
information regarding tubal patency.
Characterization of uterine anomalies
can be difficult, however, and there can
be considerable overlap in findings, no-
tably with regard to differentiation of a
septate from a bicornuate uterus (33–36).
The major limitations of the procedure
are the ability to characterize only patent
canals and the inability to evaluate the
external uterine contour adequately. HSG
also entails exposure to ionizing radiation
in these typically young women.
US imaging should be performed during the secretory phase of the menstrual
cycle, when the endometrial thickness
and echo complex are better characterized (37). Imaging should not only focus
on conventional sagittal and transverse
imaging of the pelvis but also include
orthogonal images along the long axis of
the uterus to characterize the external
uterine contour. Transabdominal US is
usually best performed with a curved
4 –1-MHz or 6 –3-MHz transducer, although US is operator dependent and
may be limited because of the patient’s
body habitus, the uterine lie, and shadowing from peristaltic bowel loops. Endovaginal US should be performed with
an 8 –5-MHz endovaginal transducer; endovaginal US has the advantage of improved spatial resolution, although at the
expense of a decreased field of view. US
has a reported pooled accuracy of approximately 90%–92% (35,37,38). Hysterosonography, with infusion of saline
into the endometrial canal, provides improved delineation of the endometrium
and internal uterine morphology; however, it shares limitations similar to those
of conventional endovaginal US and can
only help evaluate patent endometrial
canals (39). Three-dimensional US with
surface- and transparent-mode reconstructions of the uterus has reported advantages over conventional two-dimensional scanning. In experienced hands, a
sensitivity of 93% and a specificity of
100% have been achieved (40). The technique allows improved delineation of the
external uterine contour and uterine volume (12,41). Further refinement of the
technique and more universal experience
and availability of the modality should
help determine its precise role in the
evaluation of müllerian duct anomalies.
Magnetic resonance (MR) imaging has
a reported accuracy of up to 100% in the
evaluation of müllerian duct anomalies
(34,35,42). Although MR imaging is
more expensive than US, its greater accuracy makes it more trusted by many gynecologists (35). Diagnostic laparoscopy,
routinely used when HSG and US were
the only available imaging modalities, is
more expensive and invasive. MR imaging provides clear delineation of internal
and external uterine anatomy in multiple
Troiano and McCarthy
Radiology
Figure 1. (a) Sagittal inversion-recovery MR image (4500/130/150 [repetition time msec/echo
time msec/inversion time msec]) is used to determine plane for obtaining images parallel to the
long axis of the uterus. (b) Coronal single-shot fast spin-echo T2-weighted MR image (minimum
repetition time, 180-msec echo time) of the retroperitoneum obtained to assess the kidneys shows
congenitally absent left kidney.
imaging planes and, most important, reliable depiction of the external uterine
contour. Complex anomalies and secondary diagnoses such as endometriosis
can often be optimally characterized
noninvasively.
Patients are best imaged with a phasedarray MR surface coil. At our institutions,
an inversion-recovery or gradient-echo
image of the uterus in the sagittal plane is
obtained initially to determine uterine
lie. Fast spin-echo T2-weighted images
are then acquired parallel to the long axis
of the uterus to characterize the external
uterine contour and are typically obtained in an oblique transverse or a coronal plane, depending on uterine lie. It is
important to obtain the images through
the long axis of the uterus immediately
after the localizing image is acquired,
otherwise urinary bladder filling may
move the uterus such that the uterus no
longer lies in the position demonstrated
on the localizing image. If the fundal
configuration is not well delineated on
the T2-weighted images, T1-weighted images parallel to the long axis can aide in
characterization of the external contour,
owing to increased contrast between the
myometrial fundal contour and the overlying fat. T1-weighted spin-echo sequences are performed with the following parameters: 600/16 (repetition time
msec/echo time msec), 20 –24-cm field of
view, one signal acquired, 256 ⫻ 160 matrix, and 4-mm section thickness with a
1-mm gap. Fast spin-echo T2-weighted
images (5000 –7500/100 –130) are acquired with a 20 –24-cm field of view,
three to four signals acquired, 256 ⫻ 256
matrix, echo train length of 16, bandVolume 233
䡠
Number 1
width of 32 Hz, and 4-mm section thickness with a 1-mm gap. Further imaging of
the pelvis with a transverse T1-weighted
sequence and additional multiplanar fast
spin-echo T2-weighted sequences may
then be performed to fully evaluate the
cervix, vagina, and ovaries. Finally, a coronal fast spoiled-gradient-echo image or a
single-shot fast spin-echo T2-weighted
image is obtained by using the body coil,
with a large field of view to enable assessment of the kidneys (Fig 1).
CLASSIFICATION OF
MÜLLERIAN DUCT ANOMALIES
The most basic classification of müllerian
duct defects consists of agenesis and hypoplasia, defects of vertical fusion, and
defects of lateral fusion. In 1979, Buttram
and Gibbons (17) proposed a classification of müllerian duct anomalies that
was based on the degree of failure of normal development, and they separated
these anomalies into classes that demonstrate similar clinical manifestations,
treatment, and prognosis for fetal salvage. Modified in 1988 by a subcommittee of the American Fertility Society (now
the American Society of Reproductive
Medicine) (43), the classification remains
the most widely accepted schematization
and addresses uterovaginal anomalies
(Fig 2).
Class I anomalies consist of segmental
agenesis and variable degrees of uterovaginal hypoplasia. Class II anomalies are
unicornuate uteri that represent partial
or complete unilateral hypoplasia. Class
III is composed of uterus didelphys in
which duplication of the uterus results
from complete nonfusion of the müllerian ducts. Class IV anomalies are bicornuate uteri that demonstrate incomplete
fusion of the superior segments of the
uterovaginal canal. Class V anomalies are
septate uteri that represent partial or
complete nonresorption of the uterovaginal septum. Class VI anomalies are arcuate uteri that result from near complete
resorption of the septum. Class VII
anomalies comprise sequelae of in utero
DES exposure. Because of the variability
and overlap of features of associated cervical and vaginal malformations, these
changes generally are not incorporated
into the basic schematics and are reported as a subset of the primary uterine
defect. Secondary classification systems
also have been introduced that further
dissect and elaborate on the original Buttram and Gibbons schema. Toaff et al
(44) described nine subtypes of septate
and bicornuate uteri that are characterized by the presence of a communication
between two otherwise separate uterocervical cavities.
Isolated congenital anomalies of the
fallopian tubes are rare, although agenesis, hypoplasia, and segmental narrowing have been reported (18). Congenital
anomalies of the ovaries are extremely
rare. Agenesis and partial development of
the ovaries have been documented, often
in association with müllerian duct and
urinary tract anomalies. Supernumerary
ovaries and accessory ovaries contiguous
with the original ovary have been described, as have been ovaries located
above the true pelvis (18,45).
It is important to note that the American Fertility Society classification system
functions as a framework for the description of anomalies, for communication
between clinicians, and for comparison
among various therapeutic modalities.
However, the most notable inherent deficiency of the classification is related to
the description of anomalies that include
features of two or more classes. These
anomalies should be described according
to their component parts and should not
be categorized into the class that most
closely approximates the dominant feature. Underscoring the importance for
appropriate reporting is the repercussion
on clinical approach and therapeutic regimen.
Septate Uterus
The septate uterus is the most common
müllerian duct anomaly. This anomaly
composes approximately 55% of mülleMüllerian Duct Anomalies
䡠
21
Radiology
Figure 2.
Classification system of müllerian duct anomalies developed by the American Fertility Society (43).
rian duct anomalies (10,11,14) and results from partial or complete failure of
resorption of the uterovaginal septum after fusion of the paramesonephric ducts.
The septate uterus is associated with
some of the poorest reproductive outcomes. The prevalence of septa in patients who have had recurrent spontaneous abortions (usually three or more) is
well known, with reported spontaneous
abortion rates ranging from 26% to 94%
(pooled data, 65%) (3,4,11,15,16,20,46 –
48). It is difficult to assess reproductive
outcome definitively, however, because
the majority of studies to date have not
been controlled (10). The septate uterus
is also associated with the worst obstetric
outcome of the müllerian duct anomalies, with overall premature birth rates
ranging from 9% to 33% (pooled data,
20%) and fetal survival rates from 10% to
75% (pooled data, 30%) (3,4,30,46 – 49).
The length of the septum does not appear
to correlate with differences in obstetric
outcome (50). Reproductive outcome has
been shown to improve after resection of
the septum, with reported decreases in
the spontaneous abortion rate from 88%
to 5.9% after hysteroscopic metroplasty
(10,51–53).
Recurrent pregnancy loss in these patients was traditionally attributed to the
fibrous and avascular nature of the septum, despite the lack of histologic data
(14,54). This assumption was brought
into question with the advent of pelvic
MR imaging, when it was observed that
22
䡠
Radiology
䡠
October 2004
Figure 3. Uterine septum. (a) Laparoscopic image shows flat external uterine fundal contour.
(b) Hysteroscopic image shows intervening septum (arrow). (Reprinted, with permission, from
reference 43.)
the signal intensity of the septum was
usually myometrial (ie, isointense to
myometrium) (34,35). In one study (35),
septal tissue obtained from five patients
at hysteroscopic metroplasty was sent for
pathologic review, and all specimens
demonstrated smooth muscle, not fibrous tissue. In another study, in which
MR signal intensity of the septum was
evaluated along with pathologic specimens (55), all partial septa contained
myometrium, whereas patients with
complete septa had evidence of myometrium in the upper segment of the septum and fibrous tissue in the lower segment. In another prospective study (56),
septal and nonseptal tissue samples were
obtained from the posterior uterine wall at
the time of Tompkin metroplasty. Multiple
biopsies demonstrated increased amounts
of muscular tissue and less connective tissue in the septum. The authors theorized
that the decreased connective tissue may
result in poor decidualization and implantation, while increased muscular tissue
may result in increased contractility of the
tissue, thereby predisposing the patient to
spontaneous abortion. In addition to the
inherent deficiencies of the composition of
the septum, the overlying endometrium
has been shown to be defective (57). A
scanning electron microscopy comparison
of the septal endometrium to the endometrium overlying the lateral uterine wall
myometrium showed the septal endometrium to be irregular in morphology, with a
Troiano and McCarthy
Radiology
Figure 4. HSG demonstration of septate versus bicornuate uteri. (a) Acute angle of divergence between uterine horns is most suggestive of a septate
uterus (arrow). (b, c) Indeterminate angles of divergence may suggest either (b) septate uterus (arrow) or (c) bicornuate uterus (arrow). Final
diagnoses were based on subsequent MR imaging results (not shown).
Figure 5. Septate uterus. (a) HSG image shows wide divergence of opacified endometrial cavities
simulating a bicornuate configuration. (b) Corresponding coronal oblique fast spin-echo T2weighted MR image (6000/120 [effective]) demonstrates insinuated leiomyoma (long arrow)
within the septum, causing exaggerated separation of cavities. Note lateral wall myoma with
cystic degeneration (short arrow) also causing distortion of left endometrial cavity.
decrease in sensitivity to preovulatory hormonal changes (58). Morphologic narrowing of the cavity by the septum, causing a
reduction in endometrial capacity, has also
been implicated in the cause of poor outcomes (48).
There is consensus that vascularity
within the septum is abnormal. Inadequate vascularization and altered relationships between the endometrial and
myometrial vessels and nerves are
thought to be causative (14,54).
The septum, which arises in the midline fundus, is considered to be complete
when it extends to the external cervical
os. Extension of the septum to the upper
Volume 233
䡠
Number 1
vagina is seen in approximately 25% of
cases (46). A partial septum is variable in
length and may be mild or extend to the
endocervical canal proximal to the external os. On rare occasions, complete duplication of the cervix can occur with a
septum, and this anomaly is included in
the spectrum of findings of the American
Fertility Society classification of septate
uteri. The external uterine contour may
be convex, flat, or mildly (⬍1.0-cm) concave (10,34,35,42). The American Fertility Society and Toaff et al classifications
do not specify the minimal depth of fundal indentation for differentiation of a
septate from a bicornuate or a uterus di-
Figure 6. Coronal oblique reconstructed threedimensional endovaginal US image of a partial
uterine septum demonstrates mild indentation
of the uterine fundus with no intervening cleft
(short arrow) and septum separating endometrial cavities (long arrow). (Image courtesy of
Anna Lev-Toaff, MD, Thomas Jefferson University, Philadelphia, Pa.)
delphys. A cutoff of 1.0 cm was chosen
after subjective evaluation by gynecologists at the time of laparoscopy and hysteroscopy and, while noted to be arbitrary, has been found to be reliable for
differentiation from a bicornuate configuration (38,59) (Fig 3).
The configuration of the external uterine contour is crucial for the differentiation of a septate from a bicornuate
uterus, because widely different clinical
and interventional approaches are assigned to each anomaly (60). A septate
uterus is often treated with hysteroscopic
resection of the septum. Bicornuate uteri
rarely necessitate surgical intervention,
although Strassman metroplasty with
Müllerian Duct Anomalies
䡠
23
Radiology
Figure 7. Classification criteria for US differentiation of septate from bicornuate uteri. A, When apex (3) of the fundal external contour occurs
below a straight line between the tubal ostia (1, 2) or, B, 5 mm (arrow) above it, the uterus is bicornuate. C, When apex is more than 5 mm (arrow)
above the line, uterus is septate.
wedge resection of the medial aspect of
each uterine horn and subsequent unification of the two cavities may be considered in selected patients with recurrent
second- and third-trimester pregnancy
losses (61). It is important to recognize
that mild concavity of the external uterine contour should not be construed as a
“partial” bicornuate configuration, because these patients may not be given the
option of hysteroscopic metroplasty.
When evaluating the uterus following
hysteroscopic metroplasty, no residual
septum or evidence of a residual septum
measuring up to 1 cm in length is considered indicative of optimal resection
(62).
HSG of a septate uterus can be used to
evaluate the size and extent of septa
(63,64); however, the diagnostic accuracy
of HSG alone is only 55% for differentiation of septate from bicornuate uteri (36).
An angle of less than 75° between the
uterine horns is suggestive of a septate
uterus, and an angle of more than 105° is
more consistent with bicornuate uteri
(36,63). Unfortunately, the majority of
angles of divergence between the horns
fall within this range, and considerable
overlap between the two anomalies is
noted (Fig 4). In addition, the presence of
leiomyomas or adenomyosis within the
septum may cause secondary distortion
and widening of the angles of divergence
of the uterine horns (Fig 5). It has been
reported that when US is used in conjunction with HSG, the correct diagnosis
can be made in 90% of cases (10,31,
36,38).
At US, the echogenic endometrial cavities are separated at the fundus by the
intermediate echogenicity of the myometrium in all partial septa and within
24
䡠
Radiology
䡠
October 2004
Figure 8. MR images of complete uterine septum. (a) Transverse oblique fast spin-echo T2weighted image (7150/120) shows convex external uterine contour with upper myometrial
segment (short arrow) and lower fibrous segment (long arrow) extending to external uterine os.
(b) Transverse fast spin-echo T2-weighted image (6000/115) shows vertical septum extending to
upper third of vagina (arrow). (Reprinted, with permission, from reference 43.)
the fundal segments of complete septa
(Fig 6). The inferior segment of the complete septum is hypoechoic and reflects
the caudal fibrous component. The external uterine contour must demonstrate a
convex, flat, or mildly concave configuration and may best be appreciated on
transverse images of the uterus; however,
definitive characterization of the fundal
contour remains a potential limitation.
US has been reported to allow differentiation of a septate from a bicornuate
uterus if a true orthogonal view along the
long axis can be obtained. In this plane, a
line is drawn between the apices of the
endometrium at the level of the ostia. If
the fundal indentation of the external
uterine contour is below the interostial
line or less than 5 mm above the line,
the uterus is considered to be bicornuate
or didelphic. The septate uterus is defined
by a fundal indentation of more than
5 mm above the interostial line (10,38)
(Fig 7).
On MR images, the septate uterus is
generally normal in size; however, each
endometrial cavity appears smaller than
the configuration of a normal cavity. In
both partial and complete septa, the signal intensity within the fundal segment
of the septum is isointense with the confluent adjacent myometrium of the walls
of the anterior and posterior uterine
body. In a complete septum, the upper
fundal segment of the septum has signal
intensity similar to that of myometrium
Troiano and McCarthy
Radiology
presence of leiomyomas and adenomyosis within the septum, as documented on
MR images, further substantiates the
myometrial composition of the septum
(55) (Fig 9). Duplication of normal cervical zonal anatomy is seen in the rare instance of septate uteri with two cervices
(Fig 10).
An intercornual distance of less than
4.0 cm has been used to distinguish a
septate from a bicornuate uterus (34).
However, this measurement is a residuum of HSG criteria that were created to
compensate for the inability of HSG to
demonstrate the fundal contour and
should not be used as a differentiating
criterion with MR imaging. Furthermore,
the angle will be broadened by leiomyomas and/or adenomyosis that involve
the septum.
Arcuate Uterus
Figure 9. MR images of partial uterine septum. (a) Coronal oblique fast spin-echo T2-weighted
image (6000/120) shows flat external uterine contour with prominent upper myometrial component (short arrow) and small lower fibrous component (long arrow). (b) Coronal oblique fast
spin-echo T2-weighted image (6100/110) shows partial septum with insinuated leiomyomas
(arrow). (c, d) Coronal oblique fast spin-echo T2-weighted images (6000/105) show partial
septum with extensive adenomyosis (curved arrows). Note concave external uterine contour
(straight arrow).
Figure 10. MR images of complete uterine septum with cervical duplication. Fast spin-echo
T2-weighted images obtained in (a) transverse oblique (7300/120) and (b) coronal oblique
(7216/105) planes show enlarged cervical segment with two distinct cervices, each with preserved
zonal anatomy (arrows).
and contains a low-signal-intensity linear
band extending to the external cervical
os on T2-weighted images (35,55) (Fig 8).
Volume 233
䡠
Number 1
Only a small subset of partial septa demonstrates an inferior linear low-signal-intensity band on T2-weighted images. The
The arcuate uterus is characterized by a
mild indentation of the endometrium at
the uterine fundus as a result of near
complete resorption of the uterovaginal
septum. Classification has been problematic, because it remains unclear whether
this variant should be classified as a true
anomaly or as an anatomic variant of
normal. In the original Buttram and Gibbons classification, the arcuate uterus
was subclassified with the bicornuate
uterus because it “most closely resembled
a ‘mild’ form of bicornuate uterus” (17).
On revision of the classification by the
American Fertility Society, a separate
class was designated, because the arcuate
uterus can be distinguished from a bicornuate uterus on the basis of its complete
fundal unification (43). Data regarding
the reproductive outcomes of patients
with an arcuate uterus are extremely
limited and widely disparate. In small
studies, both poor and good obstetric
outcomes have been reported (65– 67),
although an arcuate configuration is generally thought to be compatible with normal-term gestation, with a quoted delivery rate of 85% (65). However, when all
extrauterine factors for infertility have
been excluded, hysteroscopic correction
may be considered in selected patients
with recurrent pregnancy loss who have
a prominent or broad configuration of
the fundal myometrium.
It has been proposed that when a ratio
of less than 10% between the height of
the fundal indentation and the distance
between the lateral apices of the horns is
calculated on the basis of HSG findings,
an adverse reproductive outcome is not
anticipated (10,63,67) (Fig 11). However,
Müllerian Duct Anomalies
䡠
25
Radiology
Figure 11. Diagram of arcuate uterus ratio. When ratio of height (H) to length
(L) is less than 10%, an adverse reproductive outcome
is not expected. (Reprinted,
with permission, from reference 63.)
a defining depth of the indentation to
differentiate an arcuate configuration
from a broad septum has not been established.
On HSG images, opacification of the
endometrial cavity demonstrates a single
uterine canal with a broad saddle-shaped
indentation of the uterine fundus (35)
(Fig 12).
On US images, a normal external uterine contour is noted, with a broad
smooth indentation on the fundal segment of the endometrium. The indentation may best be appreciated in the transverse plane with subtle, focal, superior
duplication of the endometrial echogenic complexes. No division of the uterine horns is noted (35).
On MR images, the normal external
uterine contour is maintained. The myometrial fundal indentation is smooth and
broad, and the signal intensity of this
region is isointense to normal myometrium. No low-signal-intensity fibrous
component is appreciated (35) (Fig 13).
Subtle indentation of the peripheral subserosal arcuate vessels at the level of the
fundus may be noted on MR images. The
orientation mimics the pattern seen in
the septate uterus and, hypothetically,
may indicate a degree of vascular aberrancy.
Bicornuate Uterus
The bicornuate uterus results from incomplete fusion of the uterovaginal
horns at the level of the fundus and accounts for approximately 10% of müllerian duct anomalies. Patients with a bicornuate uterus and no extrauterine
infertility issues usually have little difficulty conceiving. Spontaneous abortion
rates are reported to range from 28% to
26
䡠
Radiology
䡠
October 2004
Figure 12. Arcuate uterus. HSG image demonstrates broad fundal indentation (arrow).
35% (pooled data, 30%) (3,46 – 48). Premature birth rates range from 14% to
23% (pooled data, 20%); and fetal survival rates, from 57% to 63% (pooled
data, 60%) (3,45,47,48,68). The rates of
spontaneous abortion and premature delivery have also been reported to reflect
the degree of nonfusion of the horns.
Spontaneous abortion rates and preterm
delivery are reported to be higher in
women with a complete bicornuate
uterus than in those with a partial bicornuate uterus (3).
A bicornuate uterus consists of two
symmetric cornua that are fused caudad,
with communication of the endometrial
cavities—most often at the level of the
uterine isthmus. The intervening cleft of
the complete bicornuate uterus extends
to the internal cervical os (bicornuate
unicollis), while the cleft of a partial bicornuate configuration is of variable
length. A bicornuate bicollis uterus is associated with a duplicated cervix, although a degree of communication is
maintained between the two horns. At
least six variations of the bicornuate
uterus have been described in the literature (44). Longitudinal upper vaginal
septa are reported to coexist in 25% of
bicornuate uteri.
Surgical intervention is usually not indicated, and the length of subsequent
gestations often increases with increasing
parity. Strassman metroplasty has been
advocated in women with a history of
recurrent pregnancy loss and in whom
no other infertility issues have been identified (48,61). However, the benefits of
metroplasty have never been formally
studied in a prospective trial (46). The
bicornuate uterus has been reported to
have the highest associated prevalence—
Figure 13. Arcuate uterus. Transverse fast
spin-echo T2-weighted MR image (6166/130)
demonstrates nonspecific low signal intensity
of fundal myometrium (arrow).
38%— of cervical incompetence among
müllerian duct anomalies (69). Prophylactic placement of a cervical cerclage in
selected patients has been reported to increase fetal survival rates (70).
On HSG images, the horns of the endometrial cavity are usually widely separated with an intercornual angle greater
than 105°. Each horn has a fusiform appearance, with apices that taper and end
in a single fallopian tube. However, because the radiographic appearance has
such a large degree of overlap with that of
the septate uterus and because the external uterine contour cannot be characterized, differentiation from the septate
uterus is more often not possible (35,
63,64).
On US images, a large fundal cleft must
be documented with divergence of the
uterine horns and associated echogenic
endometrial complexes (35,38). Features
such as extreme anteflexion or retroflexion and the presence and deformity
caused by overlying leiomyomas may
prove to be confounding (Fig 14).
On MR images, the bicornuate uterus
demonstrates a cleft of at least 1.0 cm of
the external fundal uterine contour. The
horns demonstrate normal uterine zonal
anatomy. The endometrial-to-myometrial
ratio and width are normal in appearance
(35,42). Superimposed leiomyomas and
adenomyosis are well demonstrated
(Fig 15).
Uterus Didelphys
Uterus didelphys, which constitutes
approximately 5% of müllerian duct
anomalies, is the result of nearly complete failure of fusion of the müllerian
Troiano and McCarthy
Radiology
Figure 14. Bicornuate uterus. (a) Transverse US image and (b) corresponding transverse oblique
fast spin-echo T2-weighted MR image (7250/105) demonstrate external fundal cleft (straight
arrow) with wide divergence of endometrial cavities. Note leiomyoma (curved arrow) in right
lateral wall.
fusiform endometrial cavities, with no
communication between the two horns.
Each endometrial cavity ends in a solitary fallopian tube. However, if the
anomaly is associated with an obstructed
longitudinal vaginal septum, only one
cervical os may be depicted, and it may
be cannulated with the endometrial configuration mimicking a unicornuate
uterus (63,64) (Fig 16).
On US images, separate divergent uterine horns are identified, with a large fundal cleft. Endometrial cavities are uniformly separate, with no evidence of
communication. Two separate cervices
need to be documented.
MR imaging demonstrates two separate uteri with widely divergent apices,
two separate cervices, and usually an upper vaginal longitudinal septum. In each
uterus, the endometrial-to-myometrial
width and ratio are preserved, as is normal uterine zonal anatomy (34,35,42).
An obstructed unilateral vaginal septum
may cause apparent marked deformity of
the uterus according to the degree of associated hematometrocolpos (Fig 17).
Unicornuate Uterus
Figure 15. MR images of bicornuate uterus. Fast spin-echo T2-weighted images in (a) coronal
oblique (5650/105) and (b) transverse (6000/130) planes provide two examples of bicornuate
uteri and demonstrate wide divergence of uterine horns, with communication of endometrial
cavities in the lower uterine body (arrow).
ducts. Each müllerian duct develops its
own hemiuterus and cervix and demonstrates normal zonal anatomy with a minor degree of fusion at the level of the
cervices. No communication is present
between the duplicated endometrial cavities. A longitudinal vaginal septum is associated in 75% of these anomalies (71).
Longitudinal vaginal septa may be complicated by defects in vertical fusion that
result in a transverse vaginal septum and
subsequent hematometrocolpos. Spontaneous abortion rates are reported to
range from 32% to 52% (pooled data,
45%) (3,46 – 48). Premature birth rates
range from 20% to 45% (pooled data,
38%); and fetal survival rates, from 41%
to 64% (pooled data, 55%) (3,46 – 48).
Strassman metroplasty, leaving the duplicated cervix intact in an attempt to
Volume 233
䡠
Number 1
prevent cervical incompetence, is a consideration for selected patients with recurrent spontaneous abortions and premature deliveries (48,61,72). As with the
bicornuate uterus, however, the benefits
of intervention remain unclear because
no controlled trials have been performed
(46).
Nonobstructive uterus didelphys is
usually asymptomatic, while uterus didelphys with unilateral vaginal obstruction may become symptomatic at menarche and manifest as dysmenorrhea.
Endometriosis and pelvic adhesions have
an increased prevalence and are reported
to be secondary to retrograde menstrual
flow in the subset of patients with obstruction (73,74).
HSG demonstrates two separate endocervical canals that open into separate
Failure of one müllerian duct to elongate while the other develops normally
results in the unicornuate uterus and
accounts for approximately 20% of müllerian duct anomalies. A unicornuate
uterus may be isolated, manifesting in
35% of patients, although it is usually
associated with variable degrees of a rudimentary uterine horn. A noncavitary
rudimentary horn without associated endometrium is seen in 33% of cases, and
that with an endometrial segment is seen
in 32%. A cavitary rudimentary horn is
designated “communicating” if there is
communication with the endometrium
of the contralateral horn (10% of cases)
and “noncommunicating” if there is no
such communication (22% of cases)
(17,34,75). The embryologic predominance of the unicornuate uterus to be on
the right has not been explained (76).
Spontaneous abortion rates are reported
to range from 41% to 62% (pooled data,
50%) (46 – 48,76 –78). Reported premature birth rates range from 10% to 20%
(pooled data, 15%); and fetal survival
rates, from 38% to 57% (pooled data,
40%) (46 – 48,76 –78). Common obstetric
complications include abnormal fetal lie
and intrauterine growth retardation.
As compared with uterus didelphys,
the increased spontaneous abortion rate
and the decreased fetal survival rate of
the unicornuate uterus are incompletely
Müllerian Duct Anomalies
䡠
27
Radiology
understood, as is the pathogenesis of
pregnancy loss. While gestational capacity has been observed to be proportional
to uterine muscular organ mass (77), the
decrease in myometrial mass seen in a
unicornuate uterus and in the horn of a
didelphic uterus are common to both. It
has been hypothesized that inadequate
vascularization and compromised uteroplacental blood flow of the unicornuate
uterus result from the decreased vascular
contribution of the uterine and uteroovarian arteries from the abnormal side
(77).
As with uterus didelphys, the manifestation is usually incidental unless a noncommunicating rudimentary horn is
present. Dysmenorrhea with hematometra may manifest at menarche in this subgroup. Resection of the noncommunicating horn is indicated, not only for
symptomatic relief but also because ectopic pregnancy may occur in the horn
by means of transperitoneal sperm migration (79). Eighty-nine percent of pregnancies that arise in a noncommunicating uterine horn end in rupture (79). In
addition, the incidence of endometriosis
is increased in this subgroup, similar to
the case in an obstructed uterus didelphys (73,74). Resection of a communicating horn is also a consideration, because
pregnancies that develop in the rudimentary horn rarely yield viable offspring
(79,80). Surgical intervention in a rudimentary horn without associated endometrium is rarely indicated.
Renal abnormalities are more commonly associated with unicornuate
uterus than with other müllerian duct
anomalies and have been reported in
40% of these patients (8,30,48). The
anomaly is always ipsilateral to the rudimentary horn. Renal agenesis is the most
commonly reported abnormality, occurring in 67% of cases. Ectopic kidney,
horseshoe kidney, cystic renal dysplasia,
and duplicated collecting systems have
also been described (30,79,81).
On HSG images, speculum inspection
of the cervix demonstrates a small cervix
and a poorly developed contralateral vaginal fornix. After instillation of contrast
material, the endometrial cavity assumes
a fusiform shape, tapering at the apex
and draining into a solitary fallopian
tube. The uterus is generally shifted off of
midline. Filling of a small communicating rudimentary horn may be seen, although HSG cannot clearly delineate
noncavitary and noncommunicating rudimentary horns (60,62) (Fig 18).
On US images, the isolated unicornu28
䡠
Radiology
䡠
October 2004
Figure 16. Uterus didelphys. (a, b) HSG images show catheterization of two separate cervices
with opacification of two widely divergent noncommunicating endometrial cavities (arrow).
Figure 17. Uterus didelphys. (a, b) Transverse fast spin-echo T2-weighted MR images (7216/
130) show complete duplication of uterine horns (short arrows), with partial degree of fusion of
adjacent cervices (long arrows).
ate uterus appears small, although the
characteristically asymmetric ellipsoidal shape is difficult to appreciate (75).
Deviation of a small uterus to one side
of the pelvis may suggest the diagnosis.
A rudimentary horn in the presence of a
small uterus may confirm the diagnosis;
however, the hypoplastic horn may
simulate a prominent cervix and confound the findings. The identification
of a cavitary uterine horn may be difficult to differentiate from other types of
duplicated uterus. Three-dimensional
US may help further characterize the
anomaly (Fig 19).
On MR images, the unicornuate uterus
appears curved and elongated, with the
external uterine contour assuming a banana shape. Uterine volume is reduced,
and the configuration of the uterus is
asymmetric. Normal myometrial zonal
anatomy is maintained. The endometrium
may be uniformly narrow or may assume a
bullet shape, tapering at the apex. The
endometrial-to-myometrial width and
ratio are reported to be normal (34,35)
Figure 18. Unicornuate uterus. HSG image
shows fusiform configuration of opacified endometrial cavity (arrow), with opacification of
one fallopian tube.
(Fig 20). The appearance of the rudimentary horn is variable. When the endometrium is absent, the horn is of low signal
intensity, with loss of normal zonal anatomy. When the endometrium is present,
zonal anatomy may be preserved.
Troiano and McCarthy
Radiology
Figure 19. Unicornuate uterus. (a) Transverse and (b) sagittal two-dimensional endovaginal US images demonstrate uterus without
gross morphologic anomaly. (c) Three-dimensional reconstructed transverse oblique US
image shows an abnormal lenticular shape of
endometrial cavity (long arrow) with asymmetric tapering at the cornua (short arrow).
(Image courtesy of Anna Lev-Toaff, Thomas
Jefferson University, Philadelphia, Pa.)
Figure 20. Unicornuate uterus. Transverse oblique T2-weighted MR images (6000/105) show
(a) unicornuate uterus (short arrow) with no associated rudimentary horn and (b) unicornuate
uterus with rudimentary horn (long arrow) and no associated endometrium.
Complex Uterine Anomalies
Because müllerian duct defects are the
result of a spectrum of nonfusion, deficient
development, and/or defective canalization, the process may become arrested at
any point in development, especially if the
bidirectional theory of septal regression is
considered. Classification of these defects
must address the component parts of the
anomaly rather than simply assign it to the
original American Fertility Society class
Volume 233
䡠
Number 1
most resembling the anomaly. The American Fertility Society classification is just a
framework; not all anomalies will fit neatly
into one of its classes. A classic example is
the tendency to assign the designation of a
uterus didelphys to a uterus that demonstrates a more advanced degree of fusion
than the minor fusion of the cervices that
characterizes a didelphic uterus. In these
cases, each segment of the uterovaginal
complex should be analyzed separately.
Hence, the anomaly should be designated
as “bicornuate configuration of the uterine
horns with a noncommunicating septum
of the lower uterine body extending
through the cervix to the upper vagina”
(Fig 21). The clinical implications are important, because initial intervention may
result in resection of the septum, leaving
the cervices and horns intact, rather than
consideration of a Strassman metroplasty
in selected patients. Another example is to
designate a unicornuate uterus with a
prominent communicating rudimentary
horn as a bicornuate uterus with “asymmetric” size of the uterine horns (Fig 22).
Again, appropriate description of the defect helps direct treatment and is especially
important given the distinct differences in
interventions between the two anomalies.
DES-exposed Uterus
DES is a synthetic estrogen that was
introduced in 1948 and prescribed for
women experiencing recurrent spontaneous abortions, premature deliveries, and
other pregnancy complications (82). By
increasing the synthesis of placental steroidal hormones, DES was thought to decrease the frequency of pregnancy loss.
In utero exposure to DES was shown to
be associated with clear cell carcinoma of
the vagina, and use of the drug was
abruptly discontinued in 1971 (83). The
incidence of clear cell carcinoma arising
in these women was eventually found to
be 0.14 –1.4 per 1000 women exposed
(82,84). Structural anomalies of the uterine corpus, cervix, and vagina were subsequently described and shown to affect
reproductive potential (82). DES has been
shown to interfere with embryologic development of the mesenchyme of the
genital tract.
An estimated 2–3 million women received DES, exposing 1.0 –1.5 million offspring in utero (85). The adverse repercussions on reproductive potential are
estimated to continue for the next 10 –15
years (46). It should be noted that not all
women exposed to DES have reproductive problems. The amount of the drug
ingested, as well as when the drug was
taken during the gestation, has clinical
implications. If the drug was taken very
early in the first trimester or after 22
weeks gestation, structural abnormalities
are not likely to occur (86). While there is
no definitive evidence that women exposed to DES have decreased conception
rates, they are at a twofold increased risk
for spontaneous abortion, and there is a
ninefold increased risk for ectopic pregnancy (84,85). While some studies have
Müllerian Duct Anomalies
䡠
29
Radiology
failed to show associated adverse obstetric outcomes (85), the majority report an
increased risk of premature labor and
perinatal mortality (17,19,87). It should
be underscored that the characteristic
spectrum of uterine abnormalities associated with DES exposure have been reported in women without a history of
exposure to the medication (88). As such,
the morphologic changes may represent
a rare müllerian anomaly of the uterus
that may be expressed because of or induced by exposure to DES.
Sixty-nine percent of DES-exposed
women have uterine abnormalities detected on HSG images (85). A T-shaped
configuration of the endometrial cavity
is the most commonly associated abnormality, seen in 31% of exposed women
(84,85). Other uterine corpus anomalies
include a small hypoplastic uterus, constriction bands, a widened lower uterine
segment, a narrowed fundal segment of
the endometrial canal, irregular endometrial margins, and intraluminal filling defects (84,85). Anomalies of the fallopian
tube include foreshortening, sacculations, and fimbrial deformities with fimbrial stenosis (89). Cervical anomalies
occur in 44% of cases and include hypoplasia, anterior cervical ridge, cervical
collar, and pseudopolyps (19,84 – 86). An
abnormal cervical finding is associated
with abnormal uterine corpus changes in
86% of cases (84). Exposed women are
reported to be predisposed to cervical incompetence, secondary not only to structural changes but also to histologic
changes such as abnormal smooth muscle–to-collagen ratio and decreased cervical elastin (84).
Prophylactic cervical cerclage has been
advocated in selected women, especially
those with a history of second-trimester
losses and preterm births (46). Given these
morphologic changes, frequent transvaginal US examinations are often performed
throughout pregnancy to monitor cervical length. Benign findings of the vagina,
such as adenosis, occur in approximately
67% of cases (85). When present, associated uterine abnormalities are present in
82%.
On HSG images, cervical hypoplasia
and cervical stenosis may make cannula
insertion into the endocervical canal difficult. The radiographic features include a
narrowed irregular endocervical canal.
The opacified endometrial cavity appears
small, with a shortened upper uterine
segment, resulting in the characteristic T
configuration (Fig 23). Constriction
bands are often seen at the midfundal
segment, causing narrowing of intersti30
䡠
Radiology
䡠
October 2004
Figure 21. MR images of complex anomaly. (a) Transverse oblique T2-weighted image (6000/
120) shows bicornuate configuration of uterine horns (short arrows) with lower uterine body
septum extending through the cervix (long arrow). (b) Transverse fast spin-echo T2-weighted
image (6000/105) shows focal duplication of the vagina (arrow).
Figure 23. DES exposure. HSG image shows
T configuration of endometrial cavity (arrow).
Figure 22. Unicornuate uterus. Transverse
oblique T2-weighted MR image (7400/105)
shows large rudimentary horn demonstrating a
noncommunicating endometrial segment (arrow).
Figure 24. DES exposure. (a, b) Coronal oblique fast spin-echo T2-weighted MR images (6750/105)
show T configuration, constriction bands (long arrows), and cavitary narrowing (short arrows).
Troiano and McCarthy
Radiology
Figure 25. MR images of vaginal agenesis and uterine hypoplasia and agenesis. (a) Transverse fast spin-echo T2-weighted image (6000/130) shows
vaginal agenesis demonstrating complete absence of normal vaginal tissue (arrow). (b) Sagittal fast spin-echo T2-weighted image (6616/104) shows
uterine hypoplasia with a small uterine remnant (arrow) and no normal zonal anatomy. (c) Sagittal fast spin-echo T2-weighted image (6500/110)
shows uterine agenesis with no evidence of uterine tissue (arrow).
tial segments of the fallopian tubes. Fallopian tubes may be short and have an
irregular contour (63,88 –90).
On US images findings can be nonspecific, and definitive diagnosis may not be
possible (90,91). Constriction bands and
the classic T configuration are often extremely difficult to characterize. Endometrial cavity length and surface area, as
well as endometrial thickness, are notably smaller than normal. Cervical length
is also markedly shorter (91). Authors of
one case report (92) described findings
consistent with a T configuration, with
extreme narrowing of the vertical limb of
the cavity and slitlike narrowing of the
midline portion of the horizontal limb.
In addition, Doppler US studies have
shown an increased uterine artery pulsatility index in DES-exposed women,
which reflects reduced uterine perfusion
(91).
On MR images, uterine hypoplasia, the
T configuration of the endometrial cavity, and constriction bands can be demonstrated. The characteristic T configuration is best delineated with optimized
imaging parallel to the long axis of the
uterus. Constriction bands are characterized by focal thickening of the junctional
zone, which makes the endometrial cavity narrow and irregular (93) (Fig 24).
Müllerian Agenesis and Hypoplasia
Vaginal and uterine agenesis and hypoplasia result from variable degrees of
early failure of the paramesonephric
ducts to develop prior to fusion and compose approximately 5%–10% of müllerian duct anomalies. Mayer-RokitanskyKuster-Hauser syndrome is the most
Volume 233
䡠
Number 1
common manifestation: It results in
complete vaginal agenesis, with 90% of
cases associated with uterine agenesis.
The ovaries are normal in the majority of
cases. In approximately 10% of affected
women, isolated vaginal agenesis may
occur with an obstructed or small rudimentary uterus as a result of failure of
development of the sinovaginal bulb
(94). Symptoms at presentation depend
on the presence or absence of functioning endometrium. Complete agenesis
and hypoplasia without functioning endometrium manifests in puberty with
primary amenorrhea. Secondary sexual
characteristics are present, which reflects
normal ovarian function. Primary amenorrhea with severe cyclic pelvic pain may
reflect isolated vaginal agenesis with a
uterus that contains functional endometrium that is secondarily obstructed, resulting in hematometra. In these patients, nonsurgical corrective treatment
consists of the Frank method of vaginal
dilatation, while vaginal reconstruction
may be performed with a stent and a
split-thickness skin graft, often by means
of a modified Abbe-McIndoe procedure
(95). There is no potential for reproduction in patients with agenesis. In general,
little to no reproductive potential is
present for patients with hypoplasia, depending on the degree of hypoplasia and
the presence of functional endometrium.
HSG has no role in the evaluation of
müllerian agenesis and hypoplasia.
On US images, a normal uterus cannot
be identified. The ovaries often are normally situated. US is the first modality
chosen in the evaluation for these anomalies (96 –98). However, evaluation of a
uterine remnant may be difficult and is
precluded by the acoustic window and
peristalsing bowel loops draped into the
pelvis. US and MR imaging are, therefore,
often complementary.
On MR images, uterine agenesis and
hypoplasia are best characterized on sagittal images, while vaginal agenesis is best
demonstrated on transverse images. Müllerian agenesis results in no identifiable
uterus. Uterine hypoplasia demonstrates
abnormal low-signal-intensity myometrium
on T2-weighted images, with poorly delineated zonal anatomy. The endometrial
cavity and the myometrium are reduced
in size (99 –101) (Fig 25).
Vaginal Septum and Obstructed
Uterovaginal Anomalies
In addition to obstruction associated
with vaginal agenesis, obstruction from
vaginal septa that arise from defects of
vertical and lateral fusion are also seen in
the spectrum of müllerian duct anomalies. Defects of vertical fusion include a
transverse vaginal septum, which can occur anywhere along the vagina, although
it occurs most frequently at the junction
of the upper and middle third (98,102).
The septum is a membrane of fibrous
connective tissue with vascular and muscular components (98,103). Resultant hematocolpos in patients with a uterus with
functioning endometrium can cause marked
distention of the vagina. Dilatation of
the endometrium in hematometrocolpos
is usually less striking, secondary to decreased distensibility of the more muscular myometrium (98). A transverse vaginal septum can be present in any of the
müllerian duct anomalies, although it is
Müllerian Duct Anomalies
䡠
31
Radiology
most frequently seen associated with
uterus didelphys and with complex duplication anomalies (104). An imperforate hymen may mimic a low transverse
septum with associated hematocolpos;
however, imperforate hymen is not a
müllerian anomaly and should be distinguished from a transverse vaginal septum
(98). Hematometra in the absence of hematocolpos may indicate agenesis or severe hypoplasia of the cervix. The clinical
repercussions of making the correct diagnosis are important, because reconstructive surgery in cases of cervical agenesis
or hypoplasia is usually unsuccessful and
hysterectomy is often indicated (98,102).
Defects of lateral fusion may give rise to
longitudinal septa and are most often associated with septate and duplication
anomalies. When vaginal obstruction occurs, it is usually unilateral (98).
HSG has little role in the evaluation of
transverse septum and obstructed uterovaginal anomalies, given the inherent inability to catheterize the cervix.
On US images, hematometrocolpos
manifests as a distended vagina and endometrium with a variable appearance,
most often appearing as a cystic mass
with diffuse low-level echoes. A solid
appearance and a completely anechoic
collection have also been described
(99). The uterus is differentiated from
the cervix by the thick myometrial wall
and milder distention of the cavity of
the uterus, as compared with the thin
and often imperceptible wall of the vagina. Presence or absence of a patent
cervix is more difficult to document.
Distention of the endocervical canal indicates cervical patency. In cervical
agenesis, the lower uterine segment appears narrowed, with a funneled appearance, and there is the presence of
hematometra in the absence of hematocolpos (98). Visualization of a discrete attenuated transverse septum is
not always possible.
MR imaging is an important modality
in helping define the complexity of the
anomaly, especially when US is limited
because of field of view and distortion of
the normal anatomy due to hematometrocolpos. On MR images, hematometrocolpos is characterized by a distended vagina with imperceptible walls, with the
blood demonstrating increased signal intensity on T1-weighted images and variable signal intensity on T2-weighted images, owing to the variable age of the
retained blood. Hematometros is well delineated given the preserved zonal anatomy of the uterus, as are the cervix and
32
䡠
Radiology
䡠
October 2004
Figure 26. MR images of obstructed transverse vaginal septum. (a) Coronal fast spin-echo
T2-weighted (6500/120) and (b) sagittal fast spin-echo T2-weighted (7150/105) images show thin
transverse vaginal septum (arrow in a) resulting in severe hematocolpos (arrow in b) and milder
hematometra.
Figure 27. MR images of complex anomaly with obstruction. (a) Transverse fast spin-echo
T2-weighted (7400/105) and (b) coronal fast spin-echo T2-weighted (7150/100) images show
bicornuate configuration of uterine horns, with a lower uterine-cervical septum (straight arrow in
a) extending to the vagina and with a right transverse vaginal septum resulting in hematometrocolpos (straight arrow in b). Note solid mass in right ovary (curved arrow), which was found at
laparotomy to be a dysgerminoma.
endocervical canals, when present (Fig
26). The multiplanar capability of MR
imaging is crucial in helping delineate
complex anomalies with marked secondary distortion of the uterovaginal anatomy. Furthermore, secondary processes
that involve the ovaries and adnexa, such
as endometriosis, are often better characterized on MR than on US images (Fig
27).
CONCLUSION
Müllerian duct anomalies encompass a
wide spectrum of clinical and imaging
findings, and while many of the anomalies will be diagnosed initially at HSG
or two-dimensional US, further imaging
with MR and, potentially, three-dimensional US will often be required for a definitive diagnosis. MR imaging currently
is the study of choice because of its high
accuracy and detailed elaboration of
uterovaginal and ovarian anatomy. Laparoscopy and hysteroscopy are then reserved for women in whom interventional therapy is being undertaken, thus
reducing health care expenditures and
sparing women invasive diagnostic procedures.
Troiano and McCarthy
Radiology
References
1. Ashton D, Amin HK, Richart RM, Neuwirth RS. The incidence of asymptomatic uterine anomalies in women undergoing transcervical tubal sterilization.
Obstet Gynecol 1988; 72:28 –30.
2. Byrne J, Nussbaum-Blask A, Taylor WS,
et al. Prevalence of mullerian duct
anomalies detected at ultrasound. Am J
Med Genet 2000; 94:9 –12.
3. Heinonen PK, Saarikoski S, Pystynen P.
Reproductive performance of women
with uterine anomalies: an evaluation of
182 cases. Acta Obstet Gynecol Scand
1982; 61:157–162.
4. Maneschi F, Zupi E, Marconi D, Valli E,
Romanini C, Mancuso S. Hysteroscopically detected asymptomatic mullerian
anomalies: prevalence and reproductive
implications. J Reprod Med 1995; 40:
684 – 688.
5. Simon C, Martinez L, Pardo F, Tortajada
M, Pellicer A. mullerian defects in
women with normal reproductive outcome. Fertil Steril 1991; 56:1192–1193.
6. Stampe Sorensen S. Estimated prevalence of mullerian duct anomalies. Acta
Obstet Gynecol Scand 1988; 67:441–
445.
7. Stray-Pedersen B, Stray-Pedersen S. Etiologic factors and subsequent reproductive performance in 195 couples with a
prior history of habitual abortion. Am J
Obstet Gynecol 1984; 148:140 –146.
8. Rock JA, Schlaff WD. The obstetric consequences of uterovaginal anomalies.
Fertil Steril 1985; 43:681– 692.
9. Green LK, Harris RE. Uterine anomalies:
frequency of diagnosis and obstetric
complications. Obstet Gynecol 1976; 47:
427– 433.
10. Homer HA, Li TC, Cooke ID. The septate
uterus: a review of management and reproductive outcome. Fertil Steril 2000;
73:1–14.
11. Raga F, Bauset C, Remohi J, Bonilla-Musoles F, Simon C, Pellicer A. Reproductive impact of congenital mullerian
anomalies. Hum Reprod 1997; 12:2277–
2281.
12. Raga F, Bonilla-Musoles F, Blanes J, Osborne NG. Congenital mullerian anomalies: diagnostic accuracy of three-dimensional ultrasound. Fertil Steril 1996;
65:523–528.
13. Acien P. Shall we operate on mullerian
defects? incidence of mullerian defects
in fertile and infertile women. Hum Reprod 1997; 12:1372–1376.
14. Fedele L, Bianchi S. Hysteroscopic
metroplasty for septate uterus. Obstet
Gynecol Clin North Am 1995; 22:473–
489.
15. Raziel A, Arieli S, Bukovsky I, Caspi E,
Golan A. Investigation of the uterine
cavity in recurrent aborters. Fertil Steril
1994; 62:1080 –1082.
16. Clifford K, Rai R, Watson H, Regan L. An
informative protocol for the investigation of recurrent miscarriage: preliminary experience of 500 consecutive
cases. Hum Reprod 1994; 9:1328 –1332.
17. Buttram VC, Gibbons WE. Mullerian
anomalies: a proposed classification (an
analysis of 144 cases). Fertil Steril 1979;
32:40 – 46.
18. Golan A, Langer R, Bukovsky I, Caspi E.
Volume 233
䡠
Number 1
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Congenital anomalies of the mullerian
system. Fertil Steril 1989; 51:747–755.
Ansbacher R. Uterine anomalies and future pregnancies. Clin Perinatol 1983;
10:295–304.
Harger JH, Archer DF, Marchese SG, Muracca-Clemens M, Garver KL. Etiology of
recurrent pregnancy losses and outcome
of subsequent pregnancies. Obstet Gynecol 1983; 62:574 –581.
Heinonen PK, Pystynen PP. Primary infertility and uterine anomalies. Fertil
Steril 1983; 40:311–316.
Wajntraub G, Milwidsky A, Weiss D.
Prevention of premature delivery in a
unicornuate uterus by cervical cerclage.
Acta Obstet Gynecol Scand 1975;
54:497– 498.
Edwards JA, Gale RP. Camptobrachydactyly: a new autosomal dominant trait
with two probable homozygotes. Am J
Hum Genet 1972; 24:464 – 474.
Sarto GE, Simpson JL. Abnormalities of
the mullerian and wolffian duct systems. Birth Defects Orig Artic Ser 1978;
14:37–54.
Lee DM, Osathanondh R, Yeh J. Localization of Bcl-2 in the human fetal mullerian tract. Fertil Steril 1998; 70:135–
140.
Muller P, Musset R, Netter A, Solal R,
Vinourd JC, Gillet JY. Etat du haut appareil urinaire chez les porteuses de malformations uterines: etude de 133 observations. Presse Med 1967; 75:1331–
1336.
Larsen WJ. Development of the urogenital system. In: Human embryology.
New York, NY: Churchill Livingstone,
1993; 235–279.
Moore KL, Persaud TV. The urogenital
system: the development of the genital
system. In: The developing human: clinically oriented embryology. 6th ed. Philadelphia, Pa: Saunders, 1998; 303.
Speroff L, Glass RH, Kase NG. Development of the mullerian system. In: Mitchell C, ed. Clinical gynecologic endocrinology and infertility. 6th ed. Baltimore,
Md: Lippincott, Williams & Wilkins,
1998; 124.
Fedele L, Bianchi S, Agnoli B, Tozzi L,
Vignali M. Urinary tract anomalies associated with unicornuate uterus. J Urol
1996; 155:847– 848.
Yoder IC, Hall DA. Hysterosalpingography in the 1990s. AJR Am J Roentgenol
1991; 157:675– 683.
Krysiewicz S. Infertility in women: diagnostic evaluation with hysterosalpingography and other imaging techniques. AJR Am J Roentgenol 1992; 159:
253–261.
Mintz MC. MR evaluation of uterine
anomalies. AJR Am J Roentgenol 1987;
148:287–290.
Carrington BM, Hricak H, Nuruddin RN,
Secaf E, Laros RK Jr, Hill EC. mullerian
duct anomalies: MR imaging evaluation.
Radiology 1990; 176:715–720.
Pellerito JS, McCarthy SM, Doyle MB,
Glickman MG, DeCherney AH. Diagnosis of uterine anomalies: relative accuracy
of MR imaging, endovaginal ultrasound,
and hysterosalpingography. Radiology
1992; 183:795– 800.
Reuter KL, Daly DC, Cohen SM. Septate
versus bicornuate uteri: errors in imag-
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
ing diagnosis. Radiology 1989; 172:749 –
752.
Nicolini U, Bellotti M, Bonazzi B, Zamberletti D, Candiani GB. Can ultrasound
be used to screen uterine malformations? Fertil Steril 1987; 47:89 –93.
Fedele L, Ferrazzi E, Dorta M, Vercelllini
P, Candiani GB. Ultrasonography in the
differential diagnosis of “double” uteri.
Fertil Steril 1988; 50:361–364.
Goldberg JM, Falcone T, Attaran M. Sonohysterographic evaluation of uterine
anomalies noted on hysterosalpingography. Hum Reprod 1997; 12:2151–2153.
Kupesic S, Kurjak A. Ultrasound and
Doppler assessment of uterine anomalies. In: Kupesic S, de Ziegler D, eds. Ultrasound and infertility. Pearl River, NY:
Parthenon, 2000; 147–153.
Wu MH, Hsu CC, Huang KE. Detection
of congenital mullerian duct anomalies
using three-dimensional ultrasound.
J Clin Ultrasound 1997; 25:487– 492.
Fedele L, Dorta M, Brioschi D, Massari C,
Candiani GB. Magnetic resonance evaluation of double uteri. Obstet Gynecol
1989; 74:844 – 847.
The American Fertility Society classifications of adnexal adhesions, distal tubal
obstruction, tubal occlusion secondary
to tubal ligation, tubal pregnancies,
mullerian anomalies and intrauterine
adhesions. Fertil Steril 1988; 49:944 –
955.
Toaff ME, Lev-Toaff AS, Toaff R. Communicating uteri: review and classification with introduction of two previously
unreported types. Fertil Steril 1984; 41:
661– 679.
Rock JA, Parmley T, Murphy AA, Jones
HW. Malposition of the ovary associated
with uterine anomalies. Fertil Steril
1986; 45:561–563.
Propst AM, Hill JA. Anatomic factors associated with recurrent pregnancy loss.
Semin Reprod Med 2000; 18:341–350.
Buttram VC. Mullerian anomalies and
their management. Fertil Steril 1983; 40:
159 –163.
Patton PE, Novy MJ. Reproductive potential of the anomalous uterus. Semin
Reprod Endocrinol 1988; 6:217–233.
Gray SE, Roberts DK, Franklin RR. Fertility after metroplasty of the uterus. J Reprod Med 1984; 29:185–188.
Kupesic S, Kurjak A. Septate uterus: detection and prediction of obstetrical
complications by different forms of ultrasonography. J Ultrasound Med 1998;
17:631– 636.
DeCherney AH, Russell JB, Graebe RA,
Polan ML. Resectoscopic management
of mullerian fusion defects. Fertil Steril
1986; 45:726 –728.
Valle RF. Hysteroscopic treatment of
partial and complete uterine septum. Int
J Fertil Menopausal Stud 1996; 41:310 –
315.
Daly DC, Witten CA, Soto-Albors CE,
Riddick DH. Hysteroscopic metroplasty:
surgical technique and obstetric outcome. Fertil Steril 1983; 39:623– 628.
Fayez JA. Comparison between abdominal and hysteroscopic metroplasty. Obstet Gynecol 1986; 68:399 – 403.
Zreik TG, Troiano RN, Ghoussoub RA, et
al. Myometrial tissue in uterine septa.
Müllerian Duct Anomalies
䡠
33
Radiology
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
34
䡠
J Am Assoc Gynecol Laparosc 1998; 5:
155–160.
Dabirashrafi H, Bahadori M, Mohammad K, et al. Septate uterus: new idea on
the histologic features of the septum in
the abnormal uterus. Am J Obstet Gynecol 1995; 172(1 pt 1):105–107.
Candiani GB, Fedele L, Zamberletti D,
DeVirgiliis D, Carinelli S. Endometrial
patterns in malformed uteri. Acta Eur
Fertil 1983; 14:311–318.
Fedele L, Bianchi S, Marchini M, Franchi
D, Tozzi L, Dorta M. Ultrastructural aspects of endometrium in infertile women
with septate uterus. Fertil Steril 1996; 65:
750 –752.
Candiani GB, Ferrazzi E, Fedele L, Vercellini P, Dorta M. Sonographic evaluation of uterine morphology: a new scanning technique. Acta Eur Fertil 1986; 17:
345–348.
Fedele L, Dorta M, Brioschi D, Villa L,
Arcaini L, Bianchi S. Re-examination of
the anatomic indications for hysteroscopic metroplasty. Eur J Obstet Gynecol Reprod Biol 1991; 39:127–131.
Strassmann EO. Fertility and unification
of double uterus. Fertil Steril 1966; 17:
165–176.
Fedele L, Bianchi S, Marchini M, Mezzopane R, DiNola G, Tozzi L. Residual uterine septum of less than 1 cm after hysteroscopic metroplasty does not impair
reproductive outcome. Hum Reprod
1996; 11:727–729.
Ott DJ, Fayez JA, Zagoria RJ, eds. “Congenital anomalies” in hysterosalpingography: a text and atlas. 2nd ed. Baltimore, Md: Williams & Wilkins, 1998;
59 – 69.
Zanetti E, Ferrari LR, Rossi G. Classification and radiographic features of uterine
malformations: hysterosalpingographic
study. Br J Radiol 1978; 51:161–170.
Tulandi T, Arronet GH, McInnes RA. Arcuate and bicornuate uterine anomalies
and infertility. Fertil Steril 1980; 34:362–
364.
Musich JR, Behrman SJ. Obstetric outcome before and after metroplasty in
women with uterine anomalies. Obstet
Gynecol 1978; 52:63– 66.
Stampe Sorensen S. Fundal contour of
the uterine cavity in the new syndrome
of minor mullerian duct anomalies and
oligomenorrhea: a prospective controlled
study. Am J Obstet Gynecol 1983; 145:
659 – 667.
Rock JA, Jones HW. The clinical management of the double uterus. Fertil Steril
1977; 28:798 – 806.
Golan A, Langer R, Wexler S, Seceg E,
Niv D, Menachem PD. Cervical cerclage:
its role in the pregnant anomalous
uterus. Int J Fertil 1990; 35:164 –170.
Blum M. Prevention of spontaneous
abortion by cervical suture of the malformed uterus. Int Surg 1977; 62:213–
215.
Sarto GE, Simpson JL. Abnormalities of
Radiology
䡠
October 2004
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
the mullerian and wolffian duct systems. Birth Defects 1978; 14(6C):37–54.
Capraro VJ, Chuang JT, Randall CL. Improved fetal salvage after metroplasty.
Obstet Gynecol 1968; 31:97–103.
Olive DL, Henderson DY. Endometriosis
and mullerian anomalies. Obstet Gynecol 1987; 69:412– 415.
Ugur M, Turan C, Mungan T, Kuscu E,
Senoz S. Endometriosis in association
with mullerian anomalies. Gynecol Obstet Invest 1995; 40:261–264.
Brody JM, Koelliker, Frishman GN. Unicornuate uterus: imaging appearance,
associated anomalies, and clinical applications. AJR Am J Roentgenol 1998; 171:
1341–1347.
Heinonen PK. Unicornuate uterus and
rudimentary horn. Fertil Steril 1997; 68:
224 –230.
Fedele L, Zamberletti D, Vercellini P,
Dorta M, Candiani GB. Reproductive
performance of women with unicornuate uterus. Fertil Steril 1987; 47:416 –
419.
Andrews MC, Jones HW. Impaired reproductive performance of the unicornuate uterus: intrauterine growth retardation, infertility, and recurrent abortion in five cases. Am J Obstet Gynecol
1982; 144:173–176.
Rolen AC, Choquette AJ, Semmens JP.
Rudimentary uterine horn: obstetric and
gynecologic implications. Obstet Gynecol 1966; 27:806 – 813.
O’Leary JL, O’Leary JA. Rudimentary
horn pregnancy. Obstet Gynecol 1963;
22:371–373.
Rock JA, Schlaff WD. The obstetric consequences of uterovaginal anomalies.
Fertil Steril 1985; 43:681– 692.
Herbst AL, Senekjian EK, Frey KW. Abortion and pregnancy loss among diethylstilbestrol-exposed women. Semin Endocrinol 1989; 7:124 –129.
Herbst AL, Ulfelder H, Poskanzer DC. Adenocarcinoma of the vagina: association
of maternal stilbesterol therapy with tumor appearance in young women. N Engl
J Med 1971; 284:878 – 881.
Goldberg JM, Falcone T. Effect of diethylstilbestrol on reproductive function.
Fertil Steril 1999; 72:1–7.
Kaufman RH, Adam E, Binder GL, Gerthoffer E. Upper genital tract changes and
pregnancy outcome in offspring exposed in utero to diethylstilbestrol. Am J
Obstet Gynecol 1980; 137:299 –308.
Winfield AC, Wentz AC. Diethylstilbestrol exposure in utero. In: Diagnostic
imaging of infertility. 2nd ed. Baltimore,
Md: Williams & Wilkins, 1992; 85–95.
Kaufman RH, Adam E, Hatch EE, et al.
Continued follow-up of pregnancy outcomes in diethylstilbestrol-exposed offspring. Obstet Gynecol 2000; 96:483–
489.
Rennell CL. T-shaped uterus in diethylstilbestrol (DES) exposure. AJR Am J
Roentgenol 1979; 132:979 –980.
DeCherney AH, Cholst I, Naftolin F.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
Structure and function of the fallopian
tubes following exposure to diethylstilbestrol (DES) during gestation. Fertil
Steril 1981; 36:741–745.
Kipersztok S, Javitt M, Hill MC, Stillman
RJ. Comparison of magnetic resonance
imaging and transvaginal sonography
with hysterosonography in the evaluation of women exposed to diethylstilbestrol. J Reprod Med 1996; 41:347–351.
Salle B, Sergeant P, Awada A, et al. Transvaginal ultrasound studies of vascular
and morphologic changes in uteri exposed to diethylstilbestrol in utero. Hum
Reprod 1996; 11:2531–2536.
Lev-Toaff AS, Toaff ME, Friedman AC.
Endovaginal sonographic appearance of
a DES uterus. J Ultrasound Med 1990;
9:661– 664.
van Gils AP, Than RT, Falke TH, Peters
AA. Abnormalities of the uterus and cervix after diethylstilbestrol exposure: correlation of findings on MR and hysterosalpingography. AJR Am J Roentgenol
1989; 153:1235–1238.
Murray JM, Gambrell RD. Complete and
partial vaginal agenesis. J Reprod Med
1979; 22:101–105.
Lindenman E, Shepard MK, Pescovitz
OH. mullerian agenesis: an update. Obstet Gynecol 1997; 90:307–312.
Rosenberg HK, Sherman NH, Tarry WF,
Duckett JW, Snyder HM. Mayer-Rokitansky-Kuster-Hauser syndrome: US aid
to diagnosis. Radiology 1986; 161:815–
819.
Strubbe EH, Willemsen WN, Lemmens JA,
Thijn CJ, Rolland R. Mayer-RokitanskyKuster-Hauser syndrome: distinction between two forms based on excretory urographic, sonographic, and laparoscopic
findings. AJR Am J Roentgenol 1993;
160:331–334.
Blask AR, Sanders RC, Rock JA. Obstructed uterovaginal anomalies: demonstration with sonography. II. Teenagers. Radiology 1991; 179:84 – 88.
Fedele L, Dorta M, Brioschi D, Giudici
MN, Candiani GB. Magnetic resonance
imaging in Mayer-Rokitansky-KusterHauser Syndrome. Obstet Gynecol 1990;
76:593–596.
Togashi K, Nishimura K, Itoh K, et al.
Vaginal agenesis: classification by MR
imaging. Radiology 1987; 162:675– 677.
Vainright JR, Fulp CJ, Schiebler ML. MR
imaging of vaginal agenesis with hematocolpos. J Comput Assist Tomogr 1988;
12:891– 893.
Rock JA. Anomalous development of the
vagina. Semin Reprod Endocrinol 1986;
4:13–31.
Wenof M, Reyniak JV, Novendstern J,
Castadot MJ. Transverse vaginal septum. Obstet Gynecol 1979; 54:60 – 64.
Arnold BW, Gilfeather M, Woodward PJ.
Mullerian duct anomalies complicated
by obstruction: evaluation with pelvic
magnetic resonance imaging. J Womens
Imaging 2001; 3:146 –152.
Troiano and McCarthy