Kaori
Togashi,
Keiji
Yamashita,
Ikuo
Konishi,
#{149} Kazumasa
MD MD
MD
Nishimura,
#{149} Toshiya
Shibata,
#{149} Takahide
Mori,
Endometrial
made tense
ited
of endometrial
when
a cyst
on Ti-weighted
was
hypenintense
images
hypointense
sioms
(multiplicity)
the signal
was exhib-
on dithat signal
images.
regardless
intensity Surgery
of
on T2-weighted was
performed
in
293 patients, and confirmation was obtained in 354 lesions. MR imaging enabled accurate diagnosis of 77 of 86 endometrial cysts and exclusion of the diagnosis of endometrial cyst in 263 of 268 other gynecologic masses with or without internal hemorrhage. The overall diagnostic sensitivity, specificity, and accuracy were 90%, 98%, 96%, respectively. MR imaging seems to be an acceptable diagnostic test on which clinical decisions can be based
Index terms: Ovary, cysts, 852.1214 Radiology
in selecting
treatment.
Endometriosis, 852.3192 852.3117 #{149} Ovary, MR studies,
1991;
#{149} Junji
as
293
to hormonal
stimu-
hemorrhage and endometrial cysts (endometriomas) (2,3). When the suggestive clinical symptoms and pelvic signs of endometriosis are coupled with an enlarged ovary, the diagnosis of endometrial cyst is likely (2). However, carcinoma of the ovary can produce similar findings, and one cannot be certain that the adnexal masses are endometrial cysts rather than ovarian malignancies (2-4). While ultrasonography and computed tomography have played some role in the diagnosis of endometrial cyst, confident diffenemtiatiom of endometriab cysts from other admexal masses is difficult (5-9). In a previous report, we reviewed the appearance of endometrial cysts at magnetic resonance (MR) imaging and suggested the potential contribution of MR imaging in the diagnosis of this condition (10). However, 5everal reports have indicated that the MR imaging appearance of endometrial cyst is not definitive and can be seen with hemorrhagic masses as well, although MR imaging is probably more specific than other imaging techniques (11-15). The purpose of this
with internal to produce
study
was
to further
evaluate
the
potential of MR imaging in diagnosing endometrial cysts and in differentiating them from other gynecobogic
180:73-78
masses
with
(I. Konishi,
TM.),
Kyoto
University
School
of
Medicine, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto-shi, 606, Japan. Received May 30, 1990; revision requested July 17; revision received March 4, 1991; accepted March 11. Address reprint requests to K.T. C RSNA, 1991
MD MD
in a large
a clinically
series
MATERIALS A total
AND
adnexal
78 patients (with
more
than
patients
were
or without
6 months.
followed
up
medication)
There
were
for
no avail-
able data after the MR examination in three patients. Thus, 371 patients were available for inclusion in this study. MR imaging was performed with a 1.5-T superconducting magnet (Signa; GE Medical Systems, Milwaukee). Multisection spin-echo (SE) images with a short repeti-
lion time (TR) and short echo time (TE) (TR msec[FE msec = 500-600,20-25) and with a long TRITE (2,000/60-80) were acquired in the sagittat plane in all patients; axial images were obtained when needed.
The section 2.5-mm
thickness
intervals.
was 5 mm, with
Data
a 256 x 256 matrix,
were
collected
resulting
5- or with
in 0.6 x
0.6-mm formed
pixels. Signal averaging was pertwo or four times for Ti-weighted
images
and
two
times
for
T2-weighted
images.
MR images were read prospectively by one or two of the authors (K.T., K.N., I. Kimura), and official radiologic reports were tients
made with the knowledge that pahad a pelvic mass suggestive of an adnexat origin and with the knowledge of
brief clinical pitat
findings
order
We arrived trial
cyst
lowing
recorded
on the hos-
form.
at the diagnosis
when
the
criteria
lesion
of endome-
satisfied
developed
from
the fotpreviously
published data (10). We have attached importance to the presence of hyperintense signal (equal to or greater than the intensity
of fat)
on
“suggestive” cyst
that
Ti-weighted diagnosis
was
weighted
images
hyperintensity
images.
A
made
when
was
entirely
hyperintense
exhibited
(equal of urine)
sisted of multiple cysts on Ti-weighted plicity”) regardless
METHODS
of 374 consecutive
patients;
clinically
the intensity
of patients
suspected
Imaging’
on
a Ti-
homogeneous
to or greater than on T2-weighted im-
ages. A definitive diagnosis was made when a cyst that was entirely hypenintense on Ti-weighted images exhibited hypointense signal (usually mixed with hypenintense areas) on T2-weighted images (“shading”) or when the lesion con-
mass. I From the Departments of Radiology and Nuclear Medicine (K.T., K.N., I. Kimura, Y.T., KY., T.S., Y.N., J.K.) and Gynecology and Obstetrics
Tsuda, Konishi,
MR
growth of endometrial tissue the uterus (1). The most fresite of ectopic endometrial imis the ovaries, where the le-
respond
lation enlarge
on T2-weighted images (shading) or when the lesion consisted of multiple hyperintense cysts on Ti-weighted images
#{149} Yoshiaki
with
is characterized
NDOMETRIOSIS
outside quent plants
on Ti-weighted
exhibited
Nakano,
the
hypenintensity
images. A definitive made when a cyst
MD MD
Kimura,
Diagnosis
E
hyperin-
images
homogeneous
T2-weighted agnosis was
cyst
that was
#{149} Issyu
#{149} Yoshihisa
MD
Cysts:
The value of 1.5-T magnetic resonance (MR) imaging in diagnosing endometrial cysts and differentiating them from other gynecologic masses was prospectively evaluated in 374 female patients with clinically suspected adnexal masses. A suggestive diagnosis
MD MD
(aged
on T2-weighted
entirely hyperintense images (“multiof the signal intensity
images.
9-85 years; mean, 40.8 years) with a clinically suspected adnexat mass underwent MR imaging. underwent
cal follow-up.
Nineteen multiple
Surgery
of these studies
was
patients
during
performed
clini-
in
Abbreviations: SE time, TR = repetition
=
spin time.
echo,
TE
=
echo
73
la.
lb.
Figures prominent
1, 2. (1) Endometrial hyperintense signal
greater in intensity because multiplicity to the uterus (U).
With
the cyst
intensity
on
from
predominantly
with
a faint
pletety extreme than
have
area
signal. hyperintense
images
the usual
the
variable
may
signal
be
less
bright
signal
shading.
seen
attached
“Multiplicity”
importance
tensity on Ti-weighted hibiting signal intensity hematomas (intermediate weighted images) and mas
(hyperintensity
area
of hypointensity
ages)
(i6-20)
metrial
cysts
orrhagic
to
a distinct
central
on Ti-weighted
and
not
were
adnexat
im-
considered
endo-
diagnosed
as hem-
specific
signal
intensity
to be endometriat associated with
were
were
considered
in the group
and
reports
and
compared
for
surgical the
lesions
that
the MR
Surgery patients.
revealed Sixty-one
bilateral
adnexal
masses
or two
sions in an ovary. Of the 293 64 had a total of 86 endometrial
74
#{149} Radiology
a total
le-
patients, cysts
of 268 gymecobogic
in intensity
endometnial
both
another
adnexal
correctly
cyst. mass.
cysts,
diagnosed
with
than
fat on
misdiagnosed
with
as cystic
MR
with
on Ti-weighted tensity noted
functional
results) these
cyst
were
pro-
(false-negative
(Figs 5, 6). Although mine lesions satisfied
two of the criteria
diagnosis
hyperintense
on
both
(entirely Ti-
T2-
images),
emdometniab cause the suspected
cyst was not made belesions were huge and we that septations were (Fig
diagnosis
5). These
two
of
lesions
as cystic neoplasms with protein-rich fluid or with internal hemorrhage. In another case, two juxtaposed cysts, which proved an
endometrial
were
cystic
neoplasm
fluid
or internal
other
six lesions
to have
cyst
also with
and
a follic-
misdiagnosed
as
protein-rich
hemorrhage. were
internal
suspected
hemorrhage
of bevel effect”
of endometrial
made because intermediate on Ti-weighted nonmeoplastic
The
also
because
formation (21), but cyst
or the
was
not
the lesions exhibited to low signal intensity images (Fig 6). The nature of these six le-
sioms
was
suspected,
cause
five
lesions
however, consisted
beof multi-
small umibocular hematocrit effect. signal intensity
images
was
cyst
hyperim-
misdiagnosed
as other
adnexab
nine
false-negative
cases
were
of multiple cysts, their predominant signal intensity was intermediate to low on Ti-weighted images. Thirty-five endometrial cysts demonstrated both of the criteria for a definitive diagnosis-shading and multiplicity. cases
misdiagnosed
cyst,
and
masses (false-negative results). Two of these lesions had predominantly hypointense signal on both Ti- and T2weighted images. Shading was noticed in 55 endometrial cysts. (For the lesions consisting of multiple cysts, shading was considered to be present if one of those cysts demonstrated shading.) Multiplicity was noticed in 48 endometnial cysts. Although five of
the
weighted
present
the
and
The cyst shows image
similar to the intensity of fat, in 80 of the 86 emdometrial cysts. Six endometrial cysts exhibited predominantly bow signal intensity on Ti-weighted images, and these
tein-rich fluid or internal hemorrhage, five as cystic adnexal masses probably of monneoplastic mature with internal hemorrhage, and one as a hemorrhagic
Ti-weighted
The MR imaging diagnosis is definitive cysts (arrows) adhered to each other and
MR
imaging:
neoplasms
(b) SE 2,000/70.
the
77
imaging (true-positive results) (Figs 1-4). Ten diagnoses were suggestive (Fig 1) and 67 were definitive (Figs 2-4). Nine endometrial cysts were
were
(a) SE 600/20.
ple cysts and one showed a typical The predominant
an endometnial
86 endometriab
diagnosis
354 lesions in 293 patients had either
It is greater
than
had
of the presence the “hematocnit
RESULTS
239 had
and
to be
of endo-
had been clinically or radiologically pointed out. The surgeons knew imaging results before surgery.
and
cyst
ulan
metriat cysts. MR imaging records
were
implants incidentally other adnexat masses
not included
other
patients
for a suggestive
masses.
We also determined additional criteria to exclude neoplastic lesions. The diagnosis of endometnal cyst was not applied for huge lesions (eg, those protruding far above the promontorium) and for lesions with solid components or obvious septation. Tiny cysts that were hyperintense on both Ti- and T2-weighted images and located at the periphery of a mass of non-
is suggestive.
(2) Endometriat cyst in 31-year-old patient. The lesion consists of multiple hyperintense
three
to hyperin-
images, cysts extypical of acute intensity on Tisubacute hematowith
were
in
diagnosis
images.
Ten
were
2b.
MR imaging
T2-weighted
Of the
In cases with signal on
hyperintense
without
we
and
The
masses
to com-
means that the lesion consists of more than two hyperintense cysts adhering each other or neighboring structures. Since
Ti-
images,
hyperintense
hypointense shading,
lesions
can
T2-weighted
hypointense
Ti-weighted
both
patient.
than urine on the T2-weighted image. is seen. (a) SE 600/25. (b) SE 2,000170.
shading,
signal
2a.
in 46-year-old
cyst on
The
histologic
gymecobogic dometriab
were
as follows:
tumor tumor
of the
268
was
=
37),
malignant
epithelial
germ cell = 5), sarcoma (n = 3), metastasis (ii = ii), benign epithelial tumon (n = 44), benign germ cell tumor (ii = 80), gonadal stromal tumor (thecoma and fibroma) (ii = 5), nonmeoplastic cyst (n = 34), parovariam cyst (n = 8), tubal carcinoma (n = 5), pelvic inflammatory disease (n = 3), hydrosalpimx (n = 8), ectopic pregnancy (n = i), beiomyoma (n = i6), and miscellaneous (n = 8). Among these 268 lesions, the diagnosis of endometrial cyst
(n (n
diagnoses
masses other than encysts, seem in 239 patients,
correctly
malignant
excluded
with July
MR 1991
3a.
4a.
3b.
Figures 600/25.
3, 4. (3) Endometriat Q,) SE 2,000/70. One
cyst in 40-year-old of the cysts exhibits hyperintense cyst (arrowheads). (4) Endometrial and multiplicity. (a) SE 600/25. (b) SE 2,000/70.
weighted intense
images area,
(arrows).
as in the
case
However,
careful
of a large
5a.
observation
frequently
5b.
imaging in 263 (true-negative Among these 263 true-negative five unilocular cysts exhibited intensity
typical
of subacute
41-year-old
with
rhagic
adnexal
cysts
MR
were
imaging
masses
confirmed
results). cases, signal he-
as hemor-
(Fig 7). The to be hemor-
rhagic moid
functional cyst (n = 2), dercyst with internal hemorrhage (n = 1), chronic hematoma due to ectopic pregnancy (n = 1), and hemorrhagic necrosis of adnexal tumor ( n = 1). Five adnexal masses were misinterpretated as endometrial cysts at MR imaging (false-positive results) (Fig 8). The pathologic diagnosis for five filled
lesions were with chybous
mesentenic fluid, hemor-
rhagic corpus luteum oid cystadenofibroma
cyst, endometriarising within
Volume
1
180
#{149} Number
6a. patient.
The MR imaging
6b. diagnosis
was false-negative.
(a) SE 600/20.
(b) SE 2,000/70.
The le-
signal on both images. However, the diagnosis of endometriat cyst was not applied because of its huge (6) Endometrial cyst in 38-year-old patient. The MR imaging diagnosis was false-negative. (a) SE 600/25. multiple cysts with level formation (arrowheads in b). Nonneoplastic hemorrhagic mass was suspected, was not applied because of intermediate signal intensity on the Ti-weighted image.
matoma (hyperintensity with a distimct central area of hypointensity on Ti-weighted images) and were diagnosed
4b.
imaging diagnosis is definitive, as shading and multiplicity are noted. (a) SE the T2-weighted image, an inhomogeneous hypointense area is seen in a patient. The MR imaging diagnosis is definitive due to extreme shading shading may be of intermediate or even low signal intensity on T2reveals heterogeneity consisting of a hypointense area containing a hyper-
cyst.
Figures 5, 6. (5) Endometrial cyst in sion shows prominent hyperintense size and the suggestion of septation. (b) SE 2,000/60. The lesion consists of but the diagnosis of endometrial cyst
these cyst
patient. The MR faint shading. On cyst in 36-year-old Cysts with extensive
am endometrial
moma,
and
cyst,
dermoid
serous
cyst.
aging diagnosis for these suggestive of endometrial
cystade-
The
shading.
MR im-
lesions was cyst in
three
(mesenteric cyst, hemorrhagic luteum cyst, serous cystadenoma) (Fig 8) and definitive in two (endometrioid cystadenofibroma, denmoid cyst). The prominent hyperimtemsity on both Ti- and T2-weighted corpus
images
could
be easily
two of the three the reason for other suggestive emoma 3 cm in clear, as it was fluid (Fig 8). In two definitive
explained
in
suggestive cases, but hyperintensity in the case, a serous cystaddiameter, was not filled with clear yellow retrospect, one of the diagnoses was a cane-
less mistake involving images of poor quality; in this case a small denmoid cyst exhibiting chemical shift artifact was carelessly misinterpretated as an
endometrial
cyst
with
prominent
In the
other
false
definitive
case, endometrioid cystadenofibroma (no malignancy) was histologically diagnosed within an endometrial cyst 4 cm in diameter. The overall sensitivity, specificity, and accuracy of MR imaging in diagnosing
endometrial
cysts
and
separat-
ing them from other adnexal masses were 90%, 98%, and 96%, respectively. Among the 78 patients who were followed up clinically for more than 6 months with or without medication, the MR imaging diagnoses were endometrial cyst in 41 and other gynecobogic mass in 37. At the time of this report, none of the lesions presumed to be an endometrial cyst has demonstrated any signs of neoplasia such as tumor growth or positive tumor marker studies. Of the i9 patients who underwent multiple studies, mine were operated Radiology
#{149} 75
7a.
8a.
7b.
Figures
8b.
(7) Corpus luteum hematoma in 38-year-old patient. The MR imaging diagnosis was hemorrhagic adnexal mass (true-negative result). (a) SE 600/25. (b) SE 2,000/70. The lesion exhibits hyperintensity with a distinct central area of hypointensity on the Ti-weighted image and homogeneous hyperintensity on the T2-weighted image, a finding that is opposite that of shading. (8) Serous cystadenoma ifiled with clear yellow fluid in 35-year-old patient. The suggestive MR imaging diagnosis was endometrial cyst (false-positive result). (a) SE 600/25. (b) SE 2,000/60. Unilocular cyst has prominent hyperintense signal on both Ti- and T2-weighted images (arrow). There was no clear explana-
tion
7, 8.
for the signal
intensity
of the contents.
a.
b.
Figure
9.
(arrows).
Endometrial (c) SE 2,000/70
cyst in 31-year-old image obtained
patient. 6 months
on. Twelve lesions in these mime patients were considered endometrial cysts at MR imaging, and all were surgicabby confirmed. At follow-up examinatiom of four lesions, the size of the lesion was obviously decreased and the changes in the signal intensity were observed. These four lesions exhibited obvious shading on T2weighted images, although no change was observed on Ti-weighted images (Fig 9).
Endometrial tion cysts that
cysts are unique have morphologic
characteristics
and
from those We designed 76
#{149} Radiology
reten-
different
of other adnexal masses. the MR imaging criteria
(b) SE 2,000/70. is smaller, and
The cyst extensive
shows shading
for diagnosis of endometnial cyst with emphasis on findings that seem to represent these features: hyperimtense signal on Ti-weighted images, shading on T2-weighted images, and
multiplicity of hypenintense Even when very small, cysts
show
perforation extruded in the
a strong
cysts. endometrial
tendency
cyst
wall,
and
blood
of the defect
then
repeated
rupture
with
accu-
subse-
quent sealing off, and the multicentric tendency of endometrial implants, will naturally result in the unique finding
of multiple
each
other
tunes
(multiplicity).
cysts
adhered
or to neighboring The
ten-
both
images
toward perforation also exthe size of the endometnab cyst, is described as rarely larger
than an orange Endometriab
which readily of endometnial
blood
signal on (arrows).
and
on grapefruit cysts usually contents,
or tarry make cysts
(3). have
thick
aged
choco-
blood,
the surgeon (2,3). This
its characteristic
may explain the signal actenistics of endometnial
think aged
viscosity intensity charcysts-hy-
penintensity on Ti-weighted and hypointense shading weighted images. A previous report on that analyzed the signal
to
strucstrong
dency plains which
characteristic
toward
(2,3). Organization blood seals over the
prominent hyperintense has abruptly developed
late-colored
mulates anew within the cavity until perforation again takes place (2,3). This
DISCUSSION
contents
C.
(a) SE 600/25. later. The cyst
images on T2endometriosis intensity
of
endometniab foci described frequent signal intensity
a second pattern, that
of hypoimtensity
Ti-
on
both
and
T2-
July
1991
weighted predominant weighted
images (ii). We found that signal intensity on Tiimages was hyperintensity
(80 of 86 lesions) sity on both Tiages is frequently
descriptions may well explain why shading can be characteristically seen in endometrial cysts. MR imaging is known to enable reliable distinction between hemorrhagic and nonhemorrhagic adnexal
and that hypointenand T2-weighted imobserved in small
boculi situated at the periphery of the predominant cysts (Fig 3). Because an endometrial cyst is a chronic retention cyst that undergoes cyclic bleeding for many years, its predominant content is aged blood. Fresh blood resulting from rebleeding will be mixed with preexisting aged blood and will thus lose its typical signal chanacteristics
of acute
hematoma.
Rebleeding
resulting in the formation of new peripheral small loculi may retain the appearance of acute hematoma, that is, hypotensity
on
both
Ti-
and
T2-
weighted images. The blood elements in chronic stage hematoma are well known to exhibit hyperintense signal on both Ti- and T2-weighted images (16-20). On the other hand, the viscosity of a fluid and the protein concentratiom are known to influence the signal intensity of fluid (22-24). We believe that high viscosity of the contents
of the
lesion,
known
to be char-
acteristic of endometrial cysts, contributes to the mechanism of shading, the finding of a hypointense area on T2-weighted images within a cyst that is hyperintense ages. It has
certain
been
protein
significant
protein
on Ti-weighted reported that
concentration,
“cross-linking”
molecules,
macroscopically sistency (22,23).
imat a
there
is
which
is observed in contime is
not as sensitive as is T2 relaxation time, and as a result, at a certain protein concentration, there is an abrupt drop in the signal intensity on T2weighted images (22,23). Indeed, we observed
the
abrupt
development
of
shading with the shrinkage of endometrial cysts, which is associated with a decrease in free water content and an increase in protein concentration. Although we are uncertain whether the shading may be attributable to an extremely high concentration of methemogbobin or a high protein concentration, we believe that the thick viscous content is necessary to produce shading in either explamation. Gynecobogic textbooks stress the characteristic viscosity of endometrial cysts (2,3). The blood in functional cysts is said to be much more apt to
form fluid
a clot than to remain as a thick as in endometnial cysts (2). In
cystadenomas,
the
blood
mixed
with
the original contents is usually not sufficient to alter the consistency of the original fluid content (2). These Volume
180
#{149} Number
i
(21).
However,
the
ability
of
MR imaging to allow one to distinguish endometrial cysts from other adnexal masses has not been fully determined. Most authors have stated that the MR appearance of endometrial cysts is not definitive and can be seen with hemorrhagic masses as well, although MR imaging is probably more specific than other imaging techniques (li-iS). However, we believe that the MR imaging findings of other adnexal masses with internal hemorrhage are substantially different from those of endometrial cyst. Gynecobogic textbooks state that a presumptive diagnosis of endometrial cyst
can
usually
be
made
macroscopi-
cally, even before microscopic examination, on the basis of the thick chocolate-colored content and morphologic characteristics (2,3). We believe that MR imaging can reflect these unique macroscopic features of endometrial cysts, although it cannot reflect microscopic characteristics of the lesions. The signal intensity of the blood mixed with the original fluid content of the adnexal mass will be less
among
as an increase Ti relaxation
masses
than
that
of blood
on
Ti-
weighted images, as the long Ti value of the original fluid content reduces the hyperintensity of the blood. Moreover, other adnexal masses will not exhibit cyclic bleeding for many years, and as a result, they will neithen contain thick, viscous aged blood nor show repeated rupture with subsequent sealing off. Thus, they will exhibit neither shading nor multiplicity. In addition, most sions with or without
neoplastic hemorrhage
be-
have the features that suggest their neoplastic nature: huge size, septation, or a solid component. In this study, we misdiagnosed only three small
neoplastic
lesions
(serous
cyst-
adenoma, endometrioid cystadenofibroma arising in an endometrial cyst, dermoid cyst) among 354 surgically confirmed lesions. No case of malignancy was misdiagnosed as an endometrial cyst. MR imaging had a sensitivity of 90%, a specificity of 98%, and an accuracy of 96% in diagnosing endometrial cysts and in separating them from other gynecobogic masses with or without internal hemorrhage. Given the 96% accuracy, MR imaging seems to be an acceptable diagnostic test on which clinical decisions can
be
based. Of course, MR imaging will not replace the baparoscopic examination for pelvic endometriosis (11,12). We completely agree with the authors of two previous reports describing the limitation of MR imaging arising from its back of sensitivity for detection of implants (10-12). However, baparoscopy will never permit direct inspection of the contents of a mass, and occasionally it fails to reveal even rebativeby large lesions in cases with dense adhesion. Thus, MR imaging seems
to offer
additional
information
in evaluating adnexal masses suggestive of endometrial cysts. Patients with endometriosis usually present with one of three complaints: pelvic pain, adnexal mass, or infertility (2,4). Since endometriosis is a benign process that becomes quiescent with pregnancy or menopause, consideration of the natural history and severity of the disease, as well as age and reproductive status, is necessary when deciding on treatment. There are many treatment options such as observation, hormonal therapy, and conservative or radical surgery. The ultimate goal of treatment in all the protean manifestations of emdometriosis
is relief
surgery patients
of symptoms.
However,
has usually been indicated with adnexal masses, even
asymptomatic
patients.
This
in
is partly
because hormonal therapy will not cause an established endometrial cyst to regress but mainly because one cannot be certain that the adnexal masses
are
rather
than
(2,4).
With
indeed
endometriosis
an ovarian its high
malignancy
accuracy
in distin-
guishing endometrial cysts from other adnexal masses, MR imaging may neduce the need for diagnostic surgery. This will be particularly useful not only in asymptomatic patients but also in young women or women approaching menopause, in whom temporization cam be used in anticipation of spontaneous improvement with menopause or pregnancy. At the least, MR imaging will aid in selecting one of the many treatment options, from simple observation to immediate diagnostic surgery, including the choice to set back the surgery after some dissolution of adhesions with hormonal manipulation. U References 1. Sampson JA. Perforating hemorrhagic (chocolate) cyst of the ovary. Arch Surg 1921; 3:245-323. 2.
3.
Williams TJ. Endometriosis. In: Mattingly RF, Thompson JD, eds. Te Linde’s operative gynecology. 6th ed. Philadelphia: Lippincott, 1985; 257-286. Novak ER, WoodruffJD. Pelvic endo-
Radiology
#{149} 77
4.
5.
metriosis. In: Novak ER, WoodruffJD, eds. Novak’s gynecologic and obstetric pathology with clinical and endocrine relations. 8th ed. Philadelphia: Saunders, 1979; 561584. Barbieri RL, Hornstein MD. Medical therapy for endometriosis. In: Wilson EA, ed. Endometriosis. New York: Liss, 1987; iii140. Fleischer AC, James AE Jr. Millis JB,Julian C. Differential diagnosis of pelvic masses
by gray scale sonography. 6.
469-476. Walsh JW, Taylor KJW, Schwartz PE, Rosenfield
10.
11.
12.
13.
AJR 1978; 131: Wasson AT.
JFM, Gray-scale
14.
8.
9.
15.
puted
17.
of endornetriosis.
Gomori
diology
Assist Tornogr 1983; 7:257-264. RW, Walsh JW. CT in gynecologic pelvic diseases. Semin Ultrasound CT MR 1988; 9:122-142. Cornput Sawer
#{149} Radiology
Mitchell
Zimmerman hematomas:
18.
78
1986; 158:639-646. Mitchell DG, Mintz
masses:
MC,
Spritzer
MR imaging
DG.
Magnetic
ing of the adnexa. MR 1988; 9:143-157.
16.
aging
CE, et al.
Semin
JM, Grossman
20.
21.
22.
Ultrasound
CT
ing.
1985; 157:87-93.
obstructed
observation
Dillon
24.
Hayman
JM, Grossman RI, Hackney DB, HI, Zimmerman RA, Bilaniuk LT. Variable appearance of subacute intracranial hematomas on high-field spin-echo
B, Marom sinonasal
Z.
secre-
on Ti and T2 shorten1989;
172:515-520.
WP, Som PM, Fullerton
GD. pointense MR signal in chronically sated sinonasal secretions. Radiology
Hyinspis1990;
174:73-78. LA, Taber
fect of clot formation
Gomori Goldberg
cyst in a
disease.
1985; 9:183-185.
PA, Rajagopalan
Radiology
23.
HI,
RA, Bilaniuk LT. Intracranial imaging by high-field MR. Ra-
liver
Assist Tomogr
Chronically
tions: imag-
hepatic
polycystic
Unger EC, Glazer HS, Lee JKT, Ling D. MM of extracranial hematomas: preliminary observations. AJR 1986; 146:403-407. Hahn PF, Stark DD, Vici LG, Ferucci JT. Duodenal hematoma: the ring sign in MR imaging. Radiology 1986; 159:379-382. Nyberg DA, Porter BA, Olds MO, Olson DO, Anderson R, Wesby GE. MR imaging of hemorrhagic adnexal masses. J Comput Assist Tomogr 1987; 11:664-669. Som PM, Dillon WP, Fullerton GD, Zim-
merman
Radi-
resonance
MR. AJNR 1987; 8:1019-1026. Wilcox DM, WeinrebJC, Lesh P.
with
Comput 19.
observations
RI, Goldberg
of a hemorrhagic
patient
Dooms GC, Hricak H, Tscholakoff D. Adnexal structures: MR imaging. Radiology
at 1.5 T, with US and CT correlation. ology 1987; 162:319-324.
criteria. Radiology 1979; 130:391-397. Berland LL, Lawson TL, Albarelli JN, Foley WD. Ultrasonic diagnosis of ovarian and adnexal disease. Semin Ultrasound 1980; 1:17-29. Fishman EK, Scatarige JC, Saksouk FA, Rosenheim NB, Siegelmann SS. Corn-
tomography
Zawin M, McCarthy 5, Scoutt LM, Cornite F. Endornetriosis: appearance and detection at MR imaging. Radiology 1989; 171: 693-696. Arrive L, Hricak H, Martin MC. Pelvic endornetriosis: MR imaging. Radiology 1989; i71:687-692.
Adnexal
ultrasound in 204 proved gynecologic masses: accuracy and specific diagnostic 7.
Nishirnura K, Togashi K, Itoh K, et al. Endornetrial cysts of ovary: MR imaging. Radiology 1987; 162:315-318.
Ford JJ, et al. Efand retraction on
KH,
spin-echo MR images of blood: an in vitro study. AJNR 1989; 10:1155-1 158.
MR im-
July 1991