Pregnancies following ultrasound-guided drainage of tubo-ovarian abscess.
ABSTRACT To study fertility among women treated by means of ultrasound-guided drainage and antibiotics for tubo-ovarian abscess (TOA).
Retrospective cohort study.
A tertiary referral center.
One hundred women of reproductive age treated for TOA between June 1986 and July 2003.
Transvaginal ultrasound-guided drainage of TOA was performed in all patients. The procedure was repeated if a substantial amount of pus was seen using ultrasonography 2-5 days after the initial aspiration, and repeated later if necessary.
Frequency of naturally conceived pregnancies.
Twenty of 38 (52.6%; 95% CI 36.5-68.9%) women who intended to have a child achieved pregnancy naturally and became mothers. In addition, 7 (50%) of 14 women who were not on birth control on a regular basis became pregnant. No ectopic pregnancies were registered.
Ultrasound-guided drainage of TOA in combination with antibiotics seems to preserve fertility in approximately half of the patients.
- SourceAvailable from: Sherif Emil[Show abstract] [Hide abstract]
ABSTRACT: Pelvic inflammatory disease and tubo-ovarian abscesses are infections that are typically transmitted sexually. Very few reports of non-sexually transmitted tubo-ovarian abscesses in children and adults have been published. We present a case of diffuse peritonitis secondary to a tubo-ovarian abscess in a 13-year-old non-sexually active, virginal female. The causative organisms were Streptococcus viridans and Peptostreptococcus. The disease was treated with salpingectomy and adnexal debridement. Despite the absence of recognizable ovarian tissue, a left ovary was found on subsequent imaging after recovery. The case, along with a review of previous literature, is presented to raise awareness of this rare entity in girls and discuss risk factors for its occurrence.Journal of Pediatric Surgery Case Reports. 10/2013; 1(10):378–380.
Transvaginal ultrasonography of the endometrium
women with postmenopausal bleeding-A Nordic
Bengt Karlsson, MD, ~ Seth Granberg, MD," Matts Wikland, MD," Pekka Yl6stalo, MD, b
Kiserud Torvid, MD, c Karel Marsal, MD, a and Lil Valentin, MD a
G6teborg and Malm6, Sweden, Helsinki, Finland, and Trondheim, Norway
OBJECTIVE: The purpose of this study was to use transvaginal ultrasonographic measurements to find
the thickness of the endometrium below which the risk of endometrial abnormality in women with
postmenopausal bleeding is low.
STUDY DESIGN: This multicenter study was carried out at eight clinics in four Nordic countries. The
study included 1168 women with postmenopausal bleeding scheduled for curettage. Before the curettage
was performed, the thickness of the endometrium was measured with transvaginal ultrasonography. The
measurement included both endometrial layers (double-layer technique). The transvaginal
ultrasonographic measurement was compared with the histopathologic diagnosis of the curettage
RESULTS: In women with atrophic endometrium the mean endometrial thickness (+_SD) was 3.9 -+ 2.5
mm. The corresponding figures for women with endometrial cancer were 21.1 -+ 11.8 mm. No malignant
endometdum was thinner than 5 mm. In 30 women (2.8%) it was not possible to measure the thickness
of the endometrium; one of these women had endometrial cancer. The 95% confidence limit for the
probability of excluding endometrial abnormality was 5.5% when the endometrial thickness was _<4 mm
as measured by transvaginal ultrasonography.
CONCLUSION: The risk of finding pathologic endometrium at curettage when the endometrium is
<4 mm as measured by transvaginal ultrasonography is 5.5%. Thus in women with postmenopausal
bleeding and an endometrium -< 4 mm it would seem justified to refrain from curettage. (AM J OBSTET
Key words: Postmenopausal, bleeding, transvaginal ultrasonography, endometrial cancer,
In about 60% of women with postmenopausal bleed-
ing no organic cause is found.l' 2 Curettage is still used
as the "gold standard" for distinguishing between a
normal and an abnormal endometrium. The method
carries a small risk of morbidity, although this risk
From the Department of Obstetrics and Gynecology, University of
GiJteborg, ~ the Department of Obstetrics and Gynecology, University of
Helsinki, b the Department of Obstetrics and Gynecology, University of
Trondheim, c and the Department of Obstetrics and Gynecology, Uni-
versity of Lund/
Supported by Swedish Medical Research Foundation grant No.
2873, the Gdteborg Medical Society, and B&K Medical, Gentofke,
Received for publication March 8, 1994," revised October 26, 1994;
accepted November 7, 1994.
Reprint requests: Bengt Karlsson, MD, Department of Obstetrics and
Gynecology, Sahlgrenska Hospital, University of G6teborg, S-413 45
Copyright © 1995 by Mosby-Year Book, Inc.
0002-9378/95 $3.00 + 0 6/1/61852
is greater in postmenopausal women with general
health problems. During the last 20 years efforts have
focused on developing and evaluating other endome-
trial biopsy techniques that can be applied on an
outpatient basis. Although many of these biopsy tech-
niques are easy to use, well accepted, and have a high
diagnostic accuracy, they have not replaced curettage
in many countries.
Transvaginal ultrasonography is an excellent diag-
nostic method for certain gynecologic diseases? 5
Therefore it is surprising that evaluation of this diag-
nostic tool has, to date, been so limited with regard to
its diagnostic potential in women with postmenopausal
bleeding. Several studies have shown a relationship
between endometrial thickness as measured by trans-
vaginal ultrasonography and endometrial abnormality
in postmenopausal women .4-]6 However, most of these
studies include a rather small number of women at a
Volume 172, Number 5
Am J Obstet Gynecol
Karlsson et al. 1489
Table I. Endometrial thickness with regard to histopathologic diagnosis of curettage specimens (n = 1138)*
Endometrial thickness (ram)
Mean SD [ Range
*Including cervical cancer (n = 9, mean endometrial thickness 7.8 +- 3.1 mm, range 3 to 11 ram); 30 women with nonmeasurable
endometrium are not included.
Table II. Histopathologic diagnosis of curettage specimens as related to endometrial thickness as measured
by transvaginal uhrasonography
Histologic diagnosis <-4 mm 5 mm
No. 6-10 mm 11-15 mm 16-20 mm 21-25 mm >25 mm
Thirty women with nonmeasurable endometrium are not included.
single medical center, making it difficult to determine
the diagnostic value on a larger scale?' 6, 7. 9-~6
The aim of this study was to determine whether mea-
surement of endometrial thickness with transvaginal ul-
trasonography can be used to exclude endometrial ab-
normality in women with postmenopausal bleeding.
Material and methods
This study was carried out as a multicenter study in
cooperation between the Departments of Obstetrics
and Gynecology at the University of Gothenburg and
the University of Lund, Malm6, Sweden; the University
of Helsinki and the University of Turku, Finland; the
University of Trondheim, Norway; the University of
Copenhagen, Rigshospitalet, and Herlev Hospital,
Denmark; and the county hospital of Sundsvall, Swe-
den. The study was approved by the ethics committees
of all participating medical faculties.
The study was performed during 1989 through 1992
and included 1168 women with postmenopausal bleed-
ing. Their mean age was (mean _+ SD) 64 +_ 9 years
(range 41 to 91 years). MI women admitted to the
participating hospitals because of postmenopausal
bleeding were included in the study.
Awoman was considered to be postmenopausal if > 1
year had elapsed since her last menstrual bleeding. No
hormonal parameters were used to characterize the
women as postmenopausal. The study included 351
women who were receiving hormonal replacement
therapy, 165 women receiving estriol, and 186 receiving
systemic hormone replacement therapy with sequential
estrogen-progestin therapy. This means that some of
the bleeding could be related to hormone replacement
therapy. Even if this was suspected to be a reason for
the bleeding, all centers routinely performed a curet-
tage on the women in this group.
The transvaginal ultrasonography was performed ei-
ther the same day or up to 3 days before the curettage.
Different uhrasonography machines equipped with
high-frequency vaginal transducers, 5 to 7.5 MHz (B&K
Medical, Pie Data Medical, Acuson, Aloka, Siemens,
Hitachi) were used. The ultrasonographic scans were
carried out by experienced transvaginal ultrasonogra-
phy examiners. The measurements of the endometrium
were performed in the longitudinal plane, and the
thickest part was identified by scanning from cornua to
cornua. The measurement included both endometrial
layers, that is, the measurement was performed between
the two basal layers of the anterior and posterior uter-
ine walls. Thus women with an atrophic endometrium
could have a thick measurement because the cavity was
distended by a polyp or fluid. However, although such
1490 Karlsson et al.
Am J Obstet Gynecol
Table III. Probability P(x) that pathologic
diagnosis will be overlooked at a certain
endometrial thickness x (millimeters) as
measured by transvaginal ultrasonography
x (mm) P(x)% ~
P(x)% a, Estimation at a certain cutoff limit for endometrial
thickness. P(x)% b, 95% confidence limit.
Fig. 1. Ultrasonographic image of endometrium measuring
5 mm (between calipers), which turned out to be endometrial
cancer. Orientation: Longitudinal scan with anterior position
to right of figure and posterior to left. Bottom of figure is toward
findings cannot be classified as endometrial abnormal-
ity, it is the pathologic features of the uterine cavity that
has to be further evaluated. 6' 7
The histopathologic diagnosis based on the results
of the curettage was used as the "gold standard." The
curettage diagnoses were grouped into seven categories:
atrophy, hormonal effect, endometrial polyp, hy-
perplasia, endometrial carcinoma, cervical carcinoma,
and "other." The "hormonal effect" category was used
when the pathologist found a hormonal response in the
endometrium. The "other" group included hydrometra,
pyometra, and hematometra. Women whose curetted
specimens were insufficient for histopathologic diagno-
sis were included in the atrophy group. Atrophic and
hormonally influenced endometrium were considered
normal; other histologic diagnoses were considered ab-
Statistics. The mean endometrial thickness and SD
for each diagnostic category were calculated. Sensitivity,
specificity, positive and negative predictive values, and
accuracy of transvaginal ultrasonography in diagnosing
endometrial abnormality at a certain cutoff level for the
endometrial thickness were calculated. We used the
method of Bickel and Docksum ~7 to calculate the risk of
overlooking an endometrial abnormality if a certain
endometrial thickness was used as the cutoff limit.
A receiver-operator characteristic curve was drawn to
determine which cutoff level of endometrial thickness
best discriminated between normal and abnormal en-
Endometrial thickness with regard to histologic diag-
nosis is shown in Tables I and II. In 30 women (2.8%)
it was not possible to measure the thickness of the
endometrium. In this group the curettage revealed one
endometrial cancer, one hyperplasia with atypia, one
cervical cancer stage IV, five polyps, one hematometra,
19 cases of atrophy, and two "hormonal effects" on the
Women receiving hormone replacement therapy
(n = 351) with a normal endometrium had a mean
endometrial thickness of 5.2 - 3.1 mm (n = 207) and
women receiving hormone replacement therapy with an
abnormal endometrium had an endometrial thickness
of 13.2 +- 8.6 mm (n = 144). The corresponding fig-
ures for women who did not use hormone replacement
therapy (n = 787) were 4.0 -+ 2.8 mm (n = 537) and
16.5 -+ 10.7 mm (n = 250).
The probability that a pathologic diagnosis will be
overlooked at a certain endometrial thickness, as mea-
sured by transvaginal ultrasonography, is shown in Table
III. At an endometrial thickness of -< 4 mm the probabil-
ity of finding an abnormal endometrium at curettage was
3.6%, with a 95% confidence limit of 5.5%. The corre-
sponding figures for endometria -< 5 mm were 6.1% and
At a cutoff limit of 4 mm for endometrial thickness
(endometrium > 4 mm indicating pathologic features)
the sensitivity for detecting a histologically abnormal
endometrium was 96%, the specificity 68%, the positive
predictive value 61%, the negative predictive value
97%, and the accuracy 78%. The corresponding figures
at a cutofflimit of 5 mm were 94%, 78%, 69%, 96%, and
84%. Two endometrial cancers were missed when a
cutoff limit of 5 mm was used (Fig. 1). No endometrial
cancer was overlooked when a cutoff limit of 4 mm was
The sensitivity and 1-specificity for transvaginal ul-
trasonography measurements (range 1 to 72 mm) for
different cutoff limits of the endometrial thickness to
identify abnormal endometrium is illustrated by the
receiver-operator characteristic curve shown (see Statis-
tics) in Fig. 2.
Gray-scale uhrasonographic images of atrophic en-
Volume 172, Number
Am J Obstet Gynecol
Karlsson et al. 1491
,o i /
0 10 20
.c/~ 5 mm
Z 6rnm I
t 7 mill
1 - Specificity (%)
50 60 70 80 90 100
Fig. 2. Receiver-operator characteristic curve illustrating sen-
sitivity and 1-specificity for different cutoff levels of endome-
trial thickness from 1 to 72 mm. Figures under graph illustrate
endometrial thickness in millimeters.
Fig. 3. Atrophic endometrium (between white arrows) as im-
aged by transvaginal uhrasonography. Endometrial thickness
2 mm. Orientation as in Fig. 1.
polyp, and endometrial cancer are shown in Figs. 3
endometrial hyperplasia, endometrial
Postmenopausal bleeding is regarded as an indication
to perform some kind of endometrial biopsy, and curet-
tage has for decades been the method of choice for diag-
nosing endometrial abnormality. However, curettage
might not be better than other endometrial biopsy
techniques.l-2. ~9-22 Curettage and other biopsy methods
(e.g. endometrial aspiration [Vabra or Pipelle]) have a
false-negative rate of 2% to 6%.~9-24 One reason for the
inaccuracy in diagnosing endometrial abnormality by
curettage is that in approximately 60% of the curettage
procedures less than half the uterine cavity is curetted. 2°
These weaknesses of the method have to be kept in mind
when curettage is used as the "gold standard" for diag-
nosing endometrial abnormality. About 10% of women
with postmenopausal bleeding will be found to have an
endometrial carcinoma and 20% to 40% will have hyper-
plasia, endometrial polyps, or other abnormalities of the
endometrium 2. 24 Thus 50% to 60% of women with post-
menopausal bleeding will have a histopathologically be-
nign diagnosis after curettage. In these women the
bleeding is considered to be caused by a thin atrophic
endometrium or atrophic vaginal mucosa, s
Curettage is an invasive method. There is a need for
a noninvasive technique that can reduce the number of
curettages. Such a method should be easy to learn, easy
Fig. 4. Endometrial hyperplasia (between white arrows) as im-
aged by transvaginal uhrasonography. Endometrial thickness
11 mm. Orientation as in Fig. 1.
to perform, and well accepted by the patient. Further-
more, the method should make it possible to select
which women would benefit from curettage. Transvagi-
nal uhrasonography is a technique that fulfills some of
Uhrasonography has been used for studying the
endometrial growth in both premenopausal and post-
menopausal women. 6' v. 10-14 With high-frequency vaginal
transducers the endometrium can easily be studied with
regard to changes in thickness. Transvaginal ultra-
sonography has also been used for characterizing en-
dometrial carcinoma, but it has been difficult to find
any specific echo pattern characteristic of endometrial
cancer. '~5 However, transvaginal ultrasonography is an
accurate method for measuring the depth of invasion in
women with endometrial cancer. 26' 27 Because other
~492 Karlsson etal. May 1995
Am J Obstet Gynecol
Fig. 5. Endometrial polyp (between white arrows) as imaged
by transvaginal ultrasonography. Endometrial thickness 17
ram. Hysteroscopy revealed polyp but curettage showed en-
dometrial atrophy. Orientation as in Fig. 1.
Fig. 6. Endometrial cancer (between white arrows) is imaged
by transvaginal ultrasonography. Endometrial thickness 17
mm. Note thick and irregular outline of endometrium (be-
tween arrows). Orientation as in Fig. 1.
studies have failed to demonstrate any specific echo
pattern characteristic for endometrial pathologic fea-
tures and because echo pattern characterization will
necessarily be subjective, we decided to evaluate
whether it would be possible to use endometrial thick-
ness as the sole parameter for diagnosing endometrial
In this study the mean thickness of an atrophic
endometrium (double layer) as measured by transvagi-
nal ultrasonography was 3.9 mm (range 1 to 22 mm).
This is in agreement with other studies. 4' 6-~5 There were
17 women who had a thick endometrium (~ 11 mm)
and atrophic endometrium after curettage (Table II).
One possible explanation is that these women had an
endometrial polyp, which often is difficult to remove by
blind curettage. This is supported by the fact that, by
means of diagnostic hysteroscopy, we found seven en-
dometrial polyps in nine women who at an earlier
curettage had no endometrial abnormality but an en-
dometrium measuring -> 8 mm. 2s
Another interesting method to diagnose endometrial
polyps is sonohysterography, where sterile saline solu-
tion is instilled into the uterine cavity during transvagi-
The mean thickness of the endometrium in women
with endometrial cancer was 21 mm, which is in agree-
ment with previous reports. 4-6' ,-~6 In the current study
no woman with endometrial cancer had an endome-
trium thinner than 5 mm. In one publication 3° endome-
trial cancer has been reported in women with an en-
dometrial thickness -< 5 mm as measured by transvagi-
nal ultrasonography. In that study two of the 100
examined women had endometrial cancer and an en-
dometrial thickness < 5 mm. However, among nearly
2000 women examined by transvaginal ultrasonogra-
phy and reported in this and other studies, no endome-
trial cancer was found when the endometrium was
thinner than 5 ram, indicating that the risk of endome-
trial cancer at an endometrial thickness <5 mm is
indeed very small. 4.' 8-~6
The receiver:operator characteristic curve (Fig. 1)
and the calculation of the probability that a pathologic
diagnosis will be overlooked at a certain endometrial
thickness (Table III) indicate that a cutoff limit of 4 mm
for endometrial thickness as measured by transvaginal
ultrasonography is suitable for excluding endometrial
pathologic features. The calculated risk of 5.5% (upper
confidence limit) for not detecting endometrial abnor-
mality at this cutoff limit should be compared with the
false-negative rate of 2% to 6% of various biopsy tech-
niques, including curettage. ~-2' ~9-24
By accepting the 4 mm cutoff limit we could have
reduced the number of curettages by 46%. This is less
than what was suggested (60%) in our first study with
fewer women (n = 205). 6
In 30 women (2.8%) the endometrium could not be
measured. This was because of the difficulties in iden-
tifying a distinct longitudinal echo line within the
uterus. The histopathologic evaluation of the specimens
from the curettage revealed that nine of these 30
women had an endometrial abnormality. Thus an en-
dometrium that cannot be measured by transvaginal
ultrasonography in a postmenopausal woman with vagi-
nal bleeding should be considered abnormal and the
woman should be referred for curettage.
Volume 172, Number 5
Am J Obstet Gynecol
Karlsson et al, 1493
Endometrial abnormality was found in 14 women
who had an endometrium measuring _< 4 mm (Table
II). In other reports, too, endometrial polyps and hy-
perplasia have been found in endometria measuring
< 5 mm. 8' ~5
When the uttrasonographic measurement alone as
described in this study for identifying endometrial or
uterine cavity abnormality, one has to be aware of the
fact that such a measurement has a certain error that
has to be evaluated for each examiner? ~
We do not believe that stage I cervical cancer can be
detected by transvaginal ultrasonography. In this study
two women with cervical cancer had an endometrium
-< 4 mm. Thus, if transvaginal ultrasonography is to be
used clinically to select women with postmenopausal
bleeding for curettage, a cervical smear should be
performed in all women not undergoing curettage. The
cervical cytologic specimen should be taken with an
endocervical brush, since this is known to have the same
diagnostic accuracy, or even better, as curettage of the
In earlier transvaginal ultrasonography studies of
women with postmenopausal bleeding the frequency of
cervical cancer has not been reported?' 6-a6, s0 In this
study nine women (0.77%) had cervical cancer. The
current findings indicate that with 4 mm as a cutoff
limit transvaginal ultrasonography of the endometrium
in a woman with postmenopausal bleeding can exclude
endometrial abnormality with reasonable certainty, To
determine the consequences of using transvaginal ul-
trasonography to select women with postmenopausal
bleeding for curettage, a prospective Nordic study has
been started by our group.
We thank the following persons who all have contrib-
uted considerably to this study: B. Cacciatore, MD,
Helsinki, Finland; F. Jensen, J. Bang, MD, and
A. Devantier, MD, Copenhagen, Denmark; S. Eik-Nes,
MD, Trondheim, Norway; O. Piiroinen, MD, and
J. Mfienpfifi, MD, Turku, Finland; and M. Dahmoun
and M. Langeen, MD, Sundsvall, Sweden. We also
thank Erik Holmberg and Olle Hfiggstr6m for their
help with statistics.
1. Grimes D. Diagnostic dilatation and curettage: a reap-
praisal. AM J OBSTET GYNECOL 1982;142:1-6.
2. Holst J, Koskela O, yon Schoultz B. Endometrial findings
following curettage in 2018 women according to age and
indications. Ann Chit Gynaecol 1983;72:274-7.
3. Cacciatore B. Can vaginal sonography predict the status of
a tubal pregnancy: a prospective comparison between
sonography, surgical findings and serum HCG levels. Ra-
4. Goldstein SR, Nachtigall M, Snyder JR, Nachtigall L.
Endometrial assessment by vaginal uhrasonography be-
fore endometrial sampling in patients with postmeno-
pausal bleeding. AM J O~STET GYNECOL 1990;163:119-23.
5. Cacciatore B, Lehtovirta P, Wahlstr6m T, Y16stalo P. Pre-
operative sonographic evaluation of endometrial cancer.
AM J OBSTET GYNECOL 1989;160:133-7.
6. Granberg S, Wikland M, Karlsson B, Norstr6m A, Friberg
L-G. Endometrial thickness as measured by endovaginal
ultrasonography for identifying endometrial abnormality.
A~u J OBSTET GYNECOL 1991 ; 164:47-52.
7. Granberg S, Friberg L-G, Norstr6m A, Wikland M. Endo-
vaginal ultrasound scanning of women with postmeno-
pausal bleeding. J Ultrasound Med 1988;7(suppl 10):283.
8. Osmers R, V61ksen M, Schauer A. Vaginosonography for
early detection of endometrial carcinoma? Lancet 1990;
9. Fleisher A, Gordon A, Entman S, Kepple D. Transvaginal
scanning of the endometrium. J Clin Ultrasound 1990; 18:
10. Grunfeld L, Walker B, Bergh P, Sandler B, Hofmann G,
Navot D. High-resolution endovaginal ultrasonography of
the endometrium: a noninvasive test for endometrial ad-
equacy. Obstet Gynecol 1991;78:200-4.
11. Y16stalo P. Ultrasonography of endometrium. Ann Med
12. Smith P, Bakos O, Heimer G, Ulmsten U. Transvaginal
ultrasound for identifying endometrial abnormality. Acta
Obstet Gynecol Stand 1991;70:591-4.
13. Varner RE, Sparks JM, Cameron CD, Roberts L, Soong SJ.
Transvaginal sonography of the endometrium in post-
menopausal women. Obstet Gynecol 1991;78:195-9.
14. Nasri MN, Shepherd JH, Setchell ME, Lowe DG, Chard T.
Sonographic depiction of postmenopausal endometrium
with transabdorninal and transvaginal scanning. Ultra-
sound Obstet Gynecol 1991;1:279-83.
15. Weiner Z, Beck D, Rottem S, Brandes J, Thaler I. Uterine
artery flow velocity waveforms and color flow imaging in
women with perimenopausal and postmenopausal bleed-
ing. Acta Obstet Gynecol Scand 1993;72:162-6.
16. Bourne T, Campbell S, Steer C, Royston P, Whitehead M,
Collins W. Detection of endometrial cancer by transvaginal
ultrasonography with colour flow imaging and blood flow
analysis: a preliminary report. Gynecol Oncol 1991;40:
17. Bickel P, Docksum K. Mathematical statistics: basic ideas
and selected topics. Oakland, California: Holden-Day,
18. Richardson DK, Schwartz JS, Weinbaum P, Gabbe SG.
Diagnostic test in obstetrics: a method for improved evalu-
ation. AM J OBSTET GYNECOL 1985;152:613-8.
19. Vuopela S. Diagnostic accuracy and clinical applicability of
cytological and histological methods for investigating en-
dometrial carcinoma. Acta Obstet
20. Stock R, Kanbour A. Prehysterectomy curettage. Obstet
21. Stovall T, Salomon S, Ling F. Endometrial sampling prior
to hysterectomy. Obstet Gynecol 1989;73:405-9.
22. Word B, Gravlee LC, Widerman GL. The fallacy of simple
uterine curettage. Obstet Gynecol 1958;12:647-7.
23. Koonings E Moyer D, Grimes D. A randomized clinical
trial comparing Pipelle and Tis-U-Trap for endometrial
biopsy. Obstet Gynecol 1990;75:293-5.
24. MacKenzie IZ, Bibby JG. Critical assessment of dilatation
and curettage in 1029 women. Lancet 1978;2:566-8.
25. Lehtovirta E Cacciatore B, Wahlstr6m T, Y16stalo P. Ul-
trasonic assessment of endometrial cancer invasion. J Clin
26. Cacciatore B, Lehtovirta P, Wahlstr6m T, Yl~inen K,
Y16stalo P. Contribution of vaginal scanning to sono-
graphic evaluation of endometrial cancer invasion. Acta
27. Karlsson B, Norstr6m A, Granberg S, Wikland M. The use
of endovaginal ultrasound to diagnose invasion of en-
dometrial carcinoma. Ultrasound Obstet Gynecol 1992;2:
28. Karlsson B, Granberg S, Wikland M, Hellberg P. Acom-
Tsuda et al.
Am J Obstet Gynecol
parative study between transvaginal sonography and hys-
teroscopy for the detection of endometrial pathology in
women with postmenopausal bleeding. J Ultrasound Med
29. Parsons A, Lense J. Sonohysterography for endometrial
abnormalities: preliminary results. J Clin Ultrasound
30. D6rum A,.Kristensen B, Langebrekke A, S6rnes T, Skaar
O. Evaluation of endometrial thickness measured by en-
dovaginal ultrasound in women with postmenopausal
bleeding. Acta Obstet Gynecol Scand 1993;72:116-9.
31. Karlsson B, Granberg S, Ridell B, Wikland M. Endometrial
thickness as measured by transvaginal sonography: inter-
observer variation. Ultrasound Obstet Gynecol 1994;4:
32. Hoffman MS, Sterghos S, Gordy LW, Gunasekaran S,
Cavanagh D. Evaluation of the cervical canal with the
endocervical brush. Obstet Gynecol 1993;82:573-7.
Differences between Occidental and Oriental
postmenopausal women in cutoff level of endometrial
thickness for endometrial cancer screening by vaginal scan
Hiroshi Tsuda, MD, a Masami Kawabata, MD," Kazume Kawabata, MD,"
Kumio Yamamoto, MD, ~ Atsuo Hidaka, MD, ~ Naohiko Umesaki, MD, b and Sachio Ogita, MD b
The findings of this study suggest that the mean endometrial thickness and the cutoff thickness for
detection of endometrial cancer in Occidental and Oriental postmenopausal women may differ. (AM J
OBSTET GYNECOL 1995; 172:1494-5.)
Key words: Endometrial cancer, transvaginal ultrasonography, menopause, endometrial
Transvaginal ultrasonography is useful in screening
for endometrial malignancies, l It is generally recog-
nized that an endometrial thickness of 4 to 8 mm is
considered a definite cutoff thickness for detection of
endometrial malignancy by transvaginal ultrasonogra-
phy/ However, we previously suggested that the cutoff
thickness differed in Occidental and Oriental women. ~
The purpose of the current study is to clarify whether
such a difference exists.
Material and methods
One hundred fifty postmenopausal women were in-
cluded in this study. None were undergoing hormone re-
placement therapy. They were clinically evaluated by
transvaginal ultrasonography and endometrial sam-
From the Department of Obstetrics and Gynecology, Osaka City
General Hospital, ~ and the Department of Obstetrics and Gynecology,
Osaka City University Medical School. b
Received for publication January 28, 1994; revised April 25, 1994;
accepted November 18, 1994.
Reprint requests: Hiroshi Tsuda, MD, Department of Obstetrics and
Gynecology, Osaka City General Hospital, 2-13-22 Miyakojimahon-
dori, Miyakojima, Osaka, Japan.
Copyright © 1995 by Mosby-Year Book, Inc.
0002-9378/95 $3.00 + 0 6/1/62308
piing. Uhrasonography was performed with Yokogawa
Medical RT 3600 (GE Yokogawa Medical Systems, Ltd.,
Tokyo) and Aloka SSD-650 (Aloka Co., Ltd., Tokyo) ma-
chines with 5.0 MHz frequency vaginal probes. Endome-
trial assessment by transvaginal ultrasonography was
performed as described by Fleischer and Entman. 1 Pa-
tients underwent fractional curettage for histologic ex-
The 150 women enrolled were a mean of 60.6 -+ 8.0
years old. Postmenopausal bleeding was noted in 47
(31.3%). Mean endometrial thicknesses for the group of
all women, atrophic endometrium, proliferative-phase
endometrium, hyperplasia, and cancer were each
2.4 -+ 2.9, 1.3 -+ 1.2, 6.6 -+ 3.6,
5.7 -+ 3.4 mm, respectively. The relationships between
histologic findings and endometrial thickness are
shown in Table I. Of 32 women with the histopathologic
diagnosis of proliferative change (proliferative phase,
hyperplasia, or cancer), 31 had endometrial thickness
-> 3 mm. Ultrasonographic measurements of endome-
trial thickness ->3 mm exhibited 92.3% sensitivity,
73.7% specificity, 25.0% positive predictive value, and
99.0% negative predictive value for detecting endome-
7.5 +- 5.3, and