Hedvig
Hricak,
Cervical Evaluation
MD
#{149} Yoshi
C. F. Chang,
MD2
#{149} Christopher
E. Cann,
PhD
#{149} Julian
T. Parer,
MD,
PhD
Incompetence: Preliminary with MR Imaging’
The ability of magnetic resonance (MR) imaging to demonstrate cervical incompetence in nonpregnant women was investigated in a prospective study of 41 volunteers referred in random order. These included 20 patients with normal cervices, 11 with cervical incompetence of traumatic or congenital origin, and 10 with clinically small cervices due to in utero diethylstilbestrol (DES) exposure. On MR images of the normal patients, cervical length was 33.0 mm ± 1.0 (mean ± standard error of the mean) and the width of the internal cervical os was 3.3 mm ± 0.1. In patients with cervical incompetence, the cervical length did not significantly differ from those in the normal group. However, the internal cervical os was significantly wider (4.5 mm ± 0.3, P < .001), and localized irregularity of the endocervical canal was demonstrated in two patients. The MR appearance of the cervical stroma varied from normal, uniformly low signal intensity (n = 4) to uniformly (n = 3) or partially (n 4) medium-to-high signal intensity on T2-weighted images. In the patients with in utero DES exposure, the mean length of the cervical canal (22.9 mm ± 1.7) was significantly shorter than that of the normal group. The width of the internal cervical os and the MR signal intensity of the cervical stroma were normal. In summary, MR findings of a cervical length shorter than 3.1 mm (95% confidence limit), an internal cervical os wider than 4.2 mm (95% confidence limit), or abnormal signal intensity in the cervical stroma are highly suggestive of incompetent cervix and should assist in planning further therapeutic decisions.
C
incompetence occurs in 0.05%-1% of all pregnancies but may be responsible for up to 16% of middle second or early third trimester abortions (1). Three main groups of factors are considered responsible for cervical incompetence: (a) acquired, including obstetric or gynecologic trauma; (b) congenital, including congenital uterine malformation, in utero exposure to ERVICAL
diethylstilbestrol
(DES),
and
de-
creased collagen content; and (c) other, including hormonal disorders, increased prostaglandin production, and multiple gestations (1,2). Despite the use of various diagnostic imaging procedures, an obstetric history of repeated, painless, second-trimester fetal loss and the visual and digital confirmation of an open cervical os during pregnancy remain the most important basis for the diagnosis (2). However, the clinical criteria for accurate diagnosis of incompetent cervix are often difficult to define (2), and, in the primipara with congenital cervical incompetence who lacks a suggestive history or premonitory symptoms,
physical
examination
may
fail to uncover the potential risk of fetal loss. Furthermore, an objective method for recognizing a patient at risk for adverse pregnancy outcome
Index terms: Uterus, S Uterus, MR studies,
abnormalities, 854.1214
(ie, perinatal loss) in whom cervical cerclage would be beneficial is not available (3). Ultrasonography (US) has been advocated as an objective way of securing the diagnosis of incompetent cervix during pregnancy (4,5). However, artifacts can interfere with US imaging of the cervix (6). To our knowledge, there has been no reported study of the ability of US to permit discrimination between the competent and incompetent cervix in nonpregnant patients. From the very beginning of its use in clinical practice, magnetic resonance (MR) imaging has shown to be an accurate, noninvasive method for demonstrating the uterus and cervix (7). This prospective study was designed to explore the potential of MR imaging in facilitating diagnosis of cervical incompetence in nonpregnant patients. MR imaging findings were compared to findings from the patient’s obstetric history and from physical examination during pregnancy.
854.14785
PATIENTS Twenty
nonpregnant
1990;
ings
174:821-826
I From the Departments of Radiology, Box 0628, C309 (H.H., Y.C.F.C., C.E.C.) and Obstettics, Gynecology, and Reproductive Sciences (J.T.P.), University of California School of Medicine, San Francisco, CA 94143. Received October 17, 1988; revision requested December 20; final revision received November 3, 1989; accepted November 6. Address reprint requests to H.H. 2
Current
address:
Veterans Administration no, Calif. © RSNA, 1990
Department
Medical
of Radiology,
Center,
Fres-
METHODS women
with
an
obstetric history of normal cervices and 21 nonpregnant women with incompetent cervices were randomly selected from the gynecology clinic population. All 41 women had been pregnant at least once. The cervix was defined as normal or incompetent
Radiology
AND
from
based
the
patient’s
on
unequivocal
obstetric
find-
history
and/or from physical examination during pregnancy. All women were 12 months or more postpartum. The women seen consecutively in the clinic who met these criteria for inclusion and agreed to undergo MR imaging composed the study group.
Abbreviations: DES diethylstilbestrol, HSG = hysterosalpingography, LE/LC ratio of length of endometrial cavity to that of cervical canal, SD = standard deviation, SE spin echo, SEM = standard error of the mean, TE echo time, TR = repetition time.
821
#{149}
..
Table 1 Clinical Findings in 21 Women Incompetent Cervices
with *
No. of Women with Incompetent Cervices 1,;L
No. of Live Births
Obstetric History G1,P1 G2,P1 G2,P2 G2,P2 G2,P2 G3,P3 G3,P3 G3,P3
Acquired or Congenital (n = 11)*
0 1 0 1 2 2 1 0
DES Exposure (n 10)t
1
1 1
3 3
3 2 1
1 3
Figure
2
Note.-G = number of pregnancies, P number of births after 20 weeks gestation. * All 1 1 women had previous spontaneous second trimester losses and documented dilated cervices.
Seven pregnancies,
women had previous including five cases cervical dilation and two cases of ture of membranes, three patients term pregnancies, but bed rest was
complicated
t
The
20 women
with
had uncomplicated no
abnormal
women
vida, petent
n
=
prenatal niphad previous required.
normal
term
obstetric
n
at clinical
and 21
of in
utero
details
the women given
in
ranged
DES
evaluation
section
(n
exposure
of the obstetric
with incompetent Table 1. Women
in age
from
=
gynecology
history
of are
(mean,
clinic.
All
was then scheduled for the sephase of the patient’s menstrual cycle. The two participating radiologists analyzed and recorded the MR findings
cretory
patient
trician
in conference of the judgment
or of the
physical collected
from
one-way
MR image
multiple
sample
MR
tamed tune,
with an elliptical 55 X 40 cm) or with
body
coil (aperture section
0.35-T San
groups
level
imaging (MT/S
Francisco,
magnet
in
with
a
.05
was
among
magnet
Diasonics, superconducting and with (Signa;
was
(aper-
10 mm
sections. February
T2-weighted
A section
with
(In studies 1985,
images
section
T2-weighted
T2-weighted
T). Four Ti-weighted
used
for
TR
of 5 mm
gaps between
was
sagittal
images
images
were
ob-
The standard resoThe image matrix T) or 256 X 128 (at
excitations per sequences,
T2-weighted
line and
were two
used were
Image
Specifically
a
age
from
the level
South
of the GE
Medi-
su-
recorded the
level
were
endometnial
(a)
of the
internal
os,
cavity from
the
cervi-
sagittal
im-
(b)
as measured the
junction,
fundus
Used
“transi-
endocervi-
(spindle
scnibed nition
for
and
MR
of the have
Because
required
in
brief
summary supplemented MR imaging tion follows.
of
this
field
precise
defi-
of the cermeasure-
investigation,
a
MR findings
published
by additional unpublished results and anatomic correla-
images,
strength,
the
is demonstrated
as
mucosa
of the
of higher
regardless
normal
extending from Os; it surrounds
ty cervical
de-
been
the
of anatomic landmarks vix is essential for the accurate ments
the
Image
cervix
(7,9-li).
stripe external
(f) on the intensity
on Ti- and T2All measurements hard-copy image. Bein millimeter increa separate 1-mm
characteristics
uterus
area on
(myo-
the
as
manufacturer.
Criteria Analysis
niphery
os to the length
canal funnel
to be incorporated routinely available
On T2-weighted
on the
of the external
the sagittal image metrial-endometrial
the
stroma
as measured
stnoma, lower uterine
or widened and irregular, of the cervical stroma image, and (g) the signal
was designed the 5-mm scale
from
sequences.
Analysis
of
tubular
os, os as
as the disof the
or
contour
as either
of the cervical stroma weighted sequences. were taken from the cause measurements ments were required,
the
cal length
a 1.5-T
shaped) thickness sagittal
scale into
image margins
(isthmus)-endocervical tubular (cylindrical)
(e) the
leio-
cervical cervical
cervical of the
configuration
corpus
or a 95%
with
the
MR imaging
was
sections.
transverse
tamed in all patients. lution mode was used: was 256 X 256 (at 0.35 1.5
the
internal internal
low-signal-intensity (d)
and the TE was 60
thickness
i-mm
and
MR
performed
before
For
Ti-
for
unequal
Technique
Calif) 22 patients
Radiology
#{149}
by
(8).
system;
perconducting
822
with
and
analyzed
differences
T2
55 cm).
thickness
no gap between
and
were
body coil a quadrature
diameter,
2,000 or 2,200 msec
analysis of vanby a Newmantest
and
thickness was 7 mm with a 3-mm gap between sections.) With the 1.5-T unit, a quadrate-system body coil (aperture, 55 cm) was used, and images were considered Ti weighted with a TR of 600 msec and a TE of 20 or 25
data
histories
Data were
to isolate
Imaging
MR
analysis
obstetric
range
sizes
the three confidence
the
used
a
internal
Submucosal
the sagittal the inner
on
segment either
when a TR of 2,000 msec and a TE of 60 msec were used. Images were ob-
msec.
the
were
the
visible.
to the of the
between
cal canal
msec.
and
investigation,
single-factor followed
(ANOVA)
Keuls
history
zone”) width
shaped,
used
of the
correlated with physical findings. the
and without of the obste-
findings.
At the end
ance
obstetric
images were conwhen a repetition
15 or 30 msec
Standard
imaging
of each knowledge
tional the
shaped.
is also
(L)
(TR) of 500 msec and an echo time
performed
the
19 patients. system, multi-
weighted 10).
is spindle
myoma
of
The women were referred for MR imaging in the same random order that they in
canal
time
with
seen
(SE)
in
therefore
15 demarcated by the entrance of the uterme vessels (v). External os of the cervix (white arrow) is demonstrated. Endocervical
(TE)
33 years).
were
Milwaukee) 0.35-T imaging
spin-echo Ti weighted
sidered
cervices in all groups
21 to 43 years
Systems, the
With
and
cerin-
Os
tance cal
com-
waist,
measured
2; multigrato have incom-
prised two subgroups-those with acquired (history of obstetric or gynecologic trauma) or congenitally incompetent cervices (n 11) and those with a
history Further
discernible
(c)
cervices
pregnancies histories. The
(primigravida, 19) determined
cervices
of premature
Figure 1. Sagittal MR image (0.35 T, SE 2,000/60 [TR msec/TE msec]) of a normal cervix in a multigravida, multipara (G2P2) patient. co = corpus uteri and lus lower uterine segment. The level of the internal cervical os (io) is marked by the entrance of the uterine vessels and/or narrowing at the uterine waist (arrow). Note the uniform low signal intensity of the cervical stroma extending across the internal os into the lower uterine segment. Lower uterine segment and endocervical canal are tubular.
2. Sagittal MR image (1.5 T, SE 2,000/60) of an incompetent cervix. The vical stroma is thin and of higher signal tensity toward the internal os (black arrows). External shape of the uterus lacks
the
and
a low-intensity
internal to the the high-intensi-
mucus.
stroma
is often
signal
intensity,
proaches
that
transition patients.
between At histologic
of
the
The seen
that
myometrium
the
of
cervical
two varies examination,
March
peas
an
ap(7). The among the
1990
proximately
at the
level
of the
internal
os
(17). Another
landmark for the location of the internal os is a uterine contour change referred to as the “uterine waist” (Fig i).
RESULTS In the 20 multiparous women with normal mean
length
(±
nonpregnant cervices, the
standard
deviation
[SD]) of the endometrial 52.5 mm ± 5.3, the mean the the
cervical mean
canal was ratio of these
(LE/LC)
was
width
3.3 mm
Figures
Sagittal MR images (0.35 T, SE 2,000/60) of two incompetent cervices. (3) CO internal os, and EO external os. The endocervical canal is irregular and widened (long white arrow), and the stroma in the same location is indistinct. (4) The cervical stroma is of uniformly abnormal medium-to-high signal intensity, and the internal cervical os (arrows) is wide. The isthmus-endocervical canal is funnel shaped. 3, 4.
uteri,
10
Table 2 MR Image Group
of the Normal,
Normal
20
Incompetent
10
DES-exposed
10
Note-Values *
t
Length Endometrial
No. of Patients
reported
Significantly Significantly
(mm)
of Cavity
± 1 SEM,
from width from width
with
in normal in normal
and DES-exposed
Length Cervical
52.5 ± 1.2 (50.0-55.0) 52.8 ± 1.6 (49.3-56.4) 47.2 ± 1.9t (42.8-51.4)
as means
different different
Incompetent,
95%
confidence
interval
and DES-exposed group, P < .001.
also referred ment, is an
ma, and
are
better
(around
the
uterine
periphery
of the
developed
stro-
laterally
(mm) of Cervical os
for the
mean
canal
patients.
The
the
cervical
1.5.
It
thian within
P < .001.
were cervix
In the cervices
1 1 women of acquired
igin,
mean
length
myometnium
connective
and tissue
the
stroma
ent
predomiof the
cervix is usually gradual, occurring over the course of 5-10 mm (i2). The isthmus, to as the lower area of transition
uterine between
segthe
from
those
ma! cervices width of the
intensity
shaped
most stroma
extends to various degrees across the ternal cervical os (Fig 1). The isthmus sphinctenic and is functionally rather
of low
signal
internal
intensity. cervical os,
fibromuscular cervix from
junction, the corpus.
the
of the of the
location
menclature in the literature. anatomy is
The
internal uterine
After
literature
Volume
most between
174
Number
#{149}
called
os and segments
a review
(i2-i7),
the description transition
also
the
demarcates the The description
the
of the following
commonly the
3
smooth
of
the novary
used. mus-
stripe
of the
cervical
stroma inis
than anatomically determined (i6). The widths of the isthmus (14) and internal os (i4,18) change during the menstrual cycle. Because the MR appearances and the width
of the
isthmus
and
internal
os can
be similar, the level of the internal os can be recognized by the location of entrance of the uterine vessels, which traverse the broad ligament and reach the uterus ap-
with
nor-
± 0.9) was significantly greater that in the normal or DES-exposed groups (P < .001) (Table 2).
connective
The
or-
mm than
the isthmus-endocervical tubular in six patients
tion
found patients.
2). The mean cervical os (4.5
while the periphery, which is composed predominantly of smooth muscle bundles, is of medium signal intensity. For discussion of the incompetent cervix, the
tissue of the cervical stroma is important (i3), and the term cervical will refer only to the inner por-
±
(± SD) of the
of women (Table internal
internal os) (15). Below the histologic internal os, the cervical stroma is predominantly fibrous (16). On T2-weighted images, a transition can be seen as the low-
either
of
mm
with incompetent or congenital
cervix and corpus and is bounded supenionly by the upper os, termed the anatomic internal Os, and infenionly by the lower Os, termed the histologic internal os (the latter is conventionally referred to as the
anteriorly or posteriorly on T2-weighted images which is predominantly nective tissue, is of low
than
43.3
incidentally in three
(12). Therefore, the inner stroma, collagenous consignal intensity,
vessel)
in
signal intensity and from the interos (Fig 1). Nabo-
of
cysts the
tubular thickness
was
low uniformly external
was
the
were mean
stroma
extended nal to the
in parentheses.
mean
os was
cervical canal (32.5 mm ± 4.2), the length of the endometrial cavity (52.8 mm ± 4.9), and the LE/LC (1.64 ± 0.3) were not significantly differ-
located
nantly
all
3.3 ± 0.1 (3.2-3.5) 4.5 ± 0.3* (4.2-6.6) 3.0 ± 0.1 (2.7-3.3)
groups,
cle of the
the
Width Internal
33.0 ± 1.0 (31.0-35.0) 32.5 ± 13 (29.4-35.6) 22.9 ± 1.7t (19.0-26.8)
cervix is found to be composed predominantly of fibrous connective tissue (i2). The strands of smooth muscle constitute approximately 10%-i5% of the stroma, are towards
(mm)of Canal
Cervix
the
cervical
errors of the mean (SEMs). The uterme waist or narrowing of the uterine contour at the level of internal os was demonstrated in 18 patients. In the same patient group when the uterine waist was used as a point of reference for the location of the internal os, the mean (± SD) endocervical length was 3.5 mm ± 0.3. The isthmus and the
endocervical Measurements
33.0 mm ± 4.3, two lengths
± 0.2, and
internal
was of
± 0.3. Table 2 shows these as means ± standard
measurements
corpus
1.61
of the
cavity length
The
uterine
in seven
waist
of the
The endocervical in seven patients,
was
demonstrated
1 1 patients
(Fig
2).
canal was tubular spindle shaped
in
two (Fig 2), and bulky and irregular in two (Fig 3). The configuration of
in three
(Fig
canal
was
and funnel 4). In two pa-
tients in whom the endocervical canal was wide and irregular, the isthmus was tubular. The entire length of the cervical stroma was of abnormal medium-to-high signal intensity
in three patients tients, the stroma
(Fig 4). In four pawas of medium-to-
high signal intensity proximally toward the level of the internal os (Fig 5). Cervical stroma was of uniformly normal low signal intensity in four
Radiology
823
#{149}
Figures
5-7.
(5) Sagittal MR image (0.35 T, SE 2,000/60) of an incompetent cervix. The proximal cervical stroma is of medium signal inteninternal cervical os. (6) Sagittal MR image (0.35 T, SE 2,000/60) of a small cervix in a patient exposed to DES in utero. The cer(cervical length less than 2 cm), but the cervical stroma is of normal low signal intensity. Arrows internal cervical os. (7) Sagit-
sity. Arrows vix is short
=
tal MR image of high
signal
cervical
os.
patients.
those
(0.35
T, SE 2,000/60)
intensity
The
four
(black
width
patients
of the
varied
3.5 mm. The sample small for statistical
of a normal-appearing
arrows)
seen
stroma
in
from
size was analysis.
2 to too
the
than that in the normal (P < and incompetent cervix (P < groups. The length of the endo-
.001) .001) metrial cavity (47.2 ± 6.0) was also significantly shorter when compared with that in the normal group (P < .001) or the group with incompetent cervices (P < .001). The mean LE/LC (2.16 ± 0.6) was not significantly different when compared that in the normal or incompetent
cervix
group
multiple
(with
were,
cavity range
val of 31.0-35.0 50.0-55.0 mm (Fig eight
7). The of the
width rowed
(95%
(Fig
in
in two pacervix and
for cervix
that
endocervical
narrow
canal
and mean
cervical
normal
tubular thickness
stroma
was
Radiology
#{149}
and
identified
signal inwas idenwoman
vaginal
in two
adenosis
others
(Fig
7).
Hysterosalpingography (HSG) (14,18) and US (4-6) have been advocated as adjuncts to clinical evaluation of the incompetent cervix. In the
shaped these
ther
patient, HSG can show internal os and funnel-
endocervical findings were
reliable
nor
ings
diagnostic
(18). cervix
or greater
cervical
during
the
US is
US findos of 15 mm
first
trimester
(4). Hour-glass
fetal
The
isthmus
of the
membrane, thinning of the lower uterine segment, and shortening of the cervix have been reported (19) as adjunct findings. Although initial US
studies
were
recent
clinical
most
encouraging,
experience
shown
in all patients.
numerous reliable
(±
of the
demonstration
of cervical
compliance
± 0.4 and
(6). Therefore,
physical examination remains the most for diagnosis of the
SD)
during accurate
incompetent
pregnancy means
cervix.
MR
The
White
imaging
focal
arrows
length
of the
lesions
internal
offers
evaluation
a new
of the
approach
cervix.
endocervical
and
The
canal
the
external
consistently demonstrated. nal os is more difficult
can
os is
The to define,
interand
a landmark of either the entrance of the uterine vessels or the narrowing at the uterine waist can be used. The characteristics on
of the
T2-weighted
normal
images
low-signal-intensity
to the
cervical
cervix tensity
the have
cervix
include stripe
the
corre-
connective
stroma
high-signal-intensity sponding to the mucus (7,10,11).
tissue
and
the
of
inner
area correcervical mucosa and Shortening of the
and a change in the signal inof the cervical stroma during third trimester of pregnancy been
described
(9).
This
clinical
observation was further supported by laboratory studies of rhesus monkeys, in which the MR signal intensity of the cervical stroma changed in response to the injection of relaxin
(Genentech, Calif) (20).
South
corresponded
ings
of stromal
In the determined
San
These
ings
more
has
that patient physique and technical artifacts preclude
3.0 mm
were
herniation
exposure.
to the
the
pa-
women,
open
DES
adenosis.
sponding
however, to be nei-
incompetent on nonpregnant
In pregnant of an
canal; found
utero
be measured,
DISCUSSION
nonpregnant a widened
in
vaginal
inter-
demonstrated normal low signal intensity in nine patients. In the remaining patient, anterior cervical
824
was
group,
with
to represent
and
group.
and
patient
at surgery
sec-
the
for endometrial cavity) uterine waist was seen in 10 patients. The mean
of the
a multigravida
and 20 mm or greater during the ond trimester have been considered diagnostic of cervical incompetence
within
confidence
mm
in
found
lips were of heterogeneous tensity. Cervical carcinoma tified on MR images of one
tients.
6).
of the internal os was narbut not significantly different
from
The
was
were
evaluation for not performed
variations
measurements; length of the
endometrial normal
test)
however,
the length tients, the
with
Newman-Keuls
composition
There
cervix
vagina
in this
In the 10 DES-exposed women, mean length (± SD) of the cervical canal (22.9 ± 5.4) was significantly shorter
in the
imaging
to
histologic
edema
present
Francisco,
MR
all patients
referring
cian to have normal cervices mal-appearing cervices on ages. The low-signal-intensity
cal stroma patient
nal cause
was
demonstrated
and extended to the external the
width
of
find-
(20).
study,
by the
find-
from cervical the
obstetrihad
norimcervi-
MR
in each the interos. Be-
internal
os
March
1990
changes during the menstrual cycle, all the patients were imaged during the secretory phase, when the internal Os iS normally the narrowest (14,18,21). The mean measurement of the width of the internal os made at HSG study (4.43 mm ± 0.23) (22) is
incompetent cervix and in our study was seen in two of 11 patients. Another HSG finding reported with incompetent cervix is the funnelshaped isthmus-endocervical canal. This was detected in three patients in our series. In summary, MR imaging
larger
findings
than
the
one
reported
here.
The discrepancy may be due to a number of factors: (a) The HSG studies
are
performed
during
the
proliferative phase when the internal os is wider compared with its width in the secretory phase, (b) there is distention of the internal os from the injection of contrast mediurn during the HSG study, and (c) MR imaging measurements were done in the sagittal plane only. The endometrial cavity (including the endocervical canal) is not a true cayity but merely a potential slitlike space flattened in the anteroposterior diameter. Therefore, the HSG measurements, which are obtained in the transverse diameter, would differ from the measurements of MR images obtained in the anteropostenor diameter. Nevertheless, our data also do not represent the true width of the internal os because the endocervical cavity is not seen-the measurements therefore include the thickness of the mucosa. In patients with the clinical diagnosis of incompetent cervix, a number of MR features were observed, many of which were previously described clinically, were demonstrated on HSG studies, or were shown at histologic examination. Shortening of the cervical canal has been reported in the pregnant patient with an incompetent cervix (23), but, to our knowledge, the length of the cervical canal has not been investigated in the nonpregnant patient. In our patients with a history of incompetent cervices, the mean length of the cervical canal was not different from that of women in the normal group. This may explain why findings at pelvic examination are often normal in nonpregnant patients with a clinical history of incompetent cervices. The dilated internal cervical os is a known feature in patients with incompetent cervices (16) and has been previously documented clinically and at HSG. The wide internal cervical os was seen on MR images as well. The importance of the shape of the endocervical canal, whether tubular or spindle or funnel shaped, is the subject of controversy. Irregular asymmetric
widening
of the
vical canal as described however, is considered Volume
174
Number
#{149}
endocer-
by Lash suggestive 3
(24), of
combine
the
findings
de-
scribed clinically or demonstrated by HSG with a better tornographic anatomic display. Furthermore, MR imaging offers the additional information of tissue characterization. The MR signal intensity of the cervical stroma was partially or uniformly increased in seven of 1 1 patients with incompetent cervices. In our studies, histologic spedmens were not available for comparison, and therefore, the MR imaging findings
were
compared
with
known
histologic findings described in the literature (12,13). It has been postulated (12) that the connective tissue of the cervical stroma allows the cervix to act as a barrier to maintain the conceptus within the uterus. Histologic study performed on patients with incompetent cervices demonstrate fewer elastic fibers and more smooth muscle in the incompetent cervix as compared with that found in patients with normal cervices (a reflection of decreased elastin and desmosine content) (13). These histologic findings are probably responsible for the increased MR signal intensity of the cervical stroma and appear to correlate with obstetric outcome in our study. The worst pregnancy outcome occurred in the three patients in whom the cervical stroma was of uniformly high signal intensity
on
T2-weighted
images.
In
these patients, repeated middle trimester pregnancy losses occurred even with the use of cerclage. When the signal intensity of the cervical stroma was only partially high, use of cerclage produced a successful outcome in three of the four patients. In our study, the length of the cervix in DES-exposed women was significantly shorter than that in the normal group. Shortening of the exocervix (portio cervix) is a known (22) clinical finding in patients exposed to DES in utero. The width of the internal os and thickness of the cervical stroma were less than those in the normal group, but the difference was not significant. The signal intensity of the cervical stroma was normal in nine of 10 patients. Morphologic studies (22) demonstrate normal connective fibrous tissue of the cervical stroma in DES-exposed women. MR findings support the hypothesis (25)
that the cervices of patients with in utero DES exposure are different than the incompetent cervices of congenital or traumatic origin. While pregnancy outcome with cerclage has been successful (26), some investigators (25,27,28) have cautioned against the liberal use of cerciage in DES-exposed patients because of the underlying differences. In summary, MR imaging can noninvasively demonstrate the distinctive features of the cervix in the nonpregnant
patient
with
a history
of in-
competent cervix or of small due to in utero DES exposure.
cervix Be-
cause
mea-
of variation
in uterine
surements depending on the patients’ habitus and obstetric history, these data are preliminary as they are based on a relatively small number of patients in each group. The patient selection was such that only patients with clinically unequivocal diagnoses of normal cervix, incompetent cervix, or in utero DES exposure were included. A larger prospective study to demonstrate a range of findings and determine the impact of MR imaging in the management of patients with suspected cervical incompetence is needed. At present, due to the wide variation of the severity of etiologic factors of this condition, we feel that a normal MR appearance of the cervix, including its measurements and tissue characteristics, cannot preclude the diagnosis of incompetent cervix. However, when MR findings of a short cervix (cervical length less than 3.1 mm, 95% confidence limit), widened internal os (internal
os wider
than
4.2
mm,
95%
confidence limit), or stroma of abnormal signal intensity are present, they can be used as an adjunct to the clinical evaluation when diagnostic or patient management decisions in cases of possible cervical incompetence need reinforcement. U References 1.
2.
Stromme WB, Haywa EW. Intrauterine fetal death in the second trimester. Am Obstet
Gynecol
Ansari
AH,
competence:
3.
4.
5.
1963;
Reynolds a review.
J
85:223-233.
RA.
Cervical inMed 1987;
J Reprod
32:161-171. Barford DAG, Rosen MG. Cervical incompetence: diagnosis and outcome. Obstet Gynecol 1984; 64:159-163. Mahran M. The role of ultrasound in the diagnosis and management of the incompetent cervix. In: Kurjak A, ed. Recent advances in ultrasound diagnosis, vol 2. International Congress Series no. 498. Amsterdam: Elsevier, 1980; 505-514. Michaels WH, Montgomery C, Karo J, Temple J, AgerJ, Olson J. Ultrasound differentiation incompetent
of the competent cervix: prevention
Radiology
from the of pre-
825
#{149}
6.
7.
8. 9.
10.
11.
12.
13.
term delivery. Am J Obstet Gynecol 1986; 154:537-546. Confino E, Mayden KL, Giglia RV, Vermesh M, Gleicher N. Pitfalls in sonographic imaging of the incompetent uterme cervix. Acta Obstet Gynecol Scand 1986; 65:593-597. Hricak H, Alpers C, Crooks LE, Sheldon PE. Magnetic resonance imaging of the female pelvis: initial experience. AJR 1983; 141:1119-1128. Zar, Jerold H. Biostatistical analysis. Englewood Cliffs, NJ: Prentice-Hall, 1974. McCarthy SM, Stark DD, Filly RA, Callen PW, Hnicak H, Higgins CB. Obstetrical magnetic resonance imaging: maternal anatomy. Radiology 1985; 154:421-425. Togashi K, Nishimura K, Itoh K, et al. Uterine cervical cancer: assessment with high-field MR imaging. Radiology 1986; 160:431-435. Hricak H, Lacey CG, Sandles LG, Chang YCF, Winkler ML, Stern JL. Invasive cervical carcinoma: comparison of MR imaging and surgical findings. Radiology 1988; 166:623-631. Danforth DN. The distribution and functional activity of the cervical musculature. Am J Obstet Gynecol 1954; 68:1261-1271. Leppert PC, Yu SY, Keller 5, Cerreta J, Mandl I. Decreased elastic fibers and desmosine content in incompetent cervix. Am J Obstet Gynecol 1987; 157:1134-1139.
14.
15.
16.
17.
18.
19.
Mann EC, McLarn WD, Hayt DB. The physiology and clinical significance of the uterine isthmus. I. The two-stage intrauterine balloon in the diagnosis and treatment of cervical incompetence. Am J Obstet Gynecol 1961; 81:209-222. Aschoff
L.
Ueber
die
Wharton-Lierse
W,
Radiology
#{149}
ed.
Applied
Kitzmiller
term
JL.
labor.
Ultrasonography
Am
J Obstet
24.
Johnson
25.
Martin
M.
Jaffe
RB,
B,
Bergman
P. Svennerund
S.
Traction
MR.
by
ultrasound.
Congress
gy,
Bergen
Lash
The
Presented
at the
and
The
cervix:
incompetent
diagnosis
Gynecol J, Landon
28.
MT.
WC
Am
Jr. Kitchin
Fertil
Riffle
Gynecol
Mangan
MM, Egan Pregnancy
CE,
Borow
1982;
Successful cervix
in a
1979; NL,
31:217-219. Higdon
1981;
JV,
Get-
in women in utero. Am
140:194-203.
L, Burtnett-Rubin
V. Giuntoli RL, Mikuta outcome in 98 women
to diethylstilbestrol ers, and unexposed
col
III.
Gynecol
to diethylstilbestrol
Steril
EC,
di-
patient:
J Obstet JD
exposed
utero.
Samuels
of the
pregnant
pnimigravida Sandberg
os of Am
SG,
Management
of incompetent
Obstet
Nor-
treatment.
1960; 79:552-558. MB, Gabbe
management in 27.
20th
Gynecolo-
internal
and
a prospective study. 1987; 157:665-669. Nunley
of
measured
of Obstetrics
ethylstilbestrol-exposed
26.
length
1974.
AF.
Obstet Ludmir
A.
pregnancy
man CE. Pregnancy outcome exposed to diethylstilbestrol
Hricak H. Magnetic resonance imaging analysis of the effects of human relaxin on cervical ripening in rhesus monkeys (abstr). In: Proceedings of the 70th Annual Meeting of the Endocrine Society, New Orleans, 1988. 21.
Soules
31:142-146.
during
P, Mennuti
1984; Demas
1979;
dix
J PD,
CB,
P. Kivikoski
uteri
the
149:52-55. 20.
Hammond
Steril
Vaalamo
cervix
in pre-
Gynecol
AF,
Fertil 23.
anatomy
of the pelvis. New York: Springer-Verlag, 1984; 283-300. Winfield AC, Wentz AC. Diagnostic imaging of infertility. Baltimore: Williams & Wilkins, 1987; 33-34. Bartolucci L, Hill WC, Katz M, Gill PJ,
Haney
Creasman WT. Diethylstilbestrol-induced upper genital tract abnormalities.
und
notwendigkeit des begniffes isthmus uteri. Verb Deutsch Ges Path 1908; 12:314. Quoted by: Mann EC. Habitual abortion: a report, in two parts, on 160 patients. Am Obstet Gynecol 1959; 77:706-718. Mann EC. Habitual abortion: a report, in two parts, on 160 patients. Am J Obstet Gynecol 1959; 77:706-718.
for demonstrating internal os of the 2:163-167.
826
berechtigung
22.
in utero, siblings.
their Obstet
JJ. exposed mothGyne-
59:315-319.
test
incompetence of the cervix. Int J Fertil 1957;
March
1990