Upload
mariaa-endahh
View
217
Download
0
Embed Size (px)
Citation preview
8/3/2019 Ad en Om Yo Sis
1/6
Adenomyosis and junctional zone changes in patients with endometriosis
S.B. Larsena,*, E. Lundorfb, A. Forman a, M. Dueholm a
aDepartment of Gynaecology and Obstetrics, Aarhus University Hospital, Skejby, DenmarkbDepartment of Diagnostic Imaging, Aarhus University Hospital, Skejby, Denmark
1. Introduction
Adenomyosis and endometriosis are both characterised by
ectopic growth of endometrium-like or endometrium-derived
tissue [1,2] and might be causally related. Leyendecker suggested
that abnormal function of the inner smoothmuscle of the uterus,
the archimetra, or the junctional zone (JZ), could represent a
common pathogenetic factor [3]. Alterations in the JZ thickness
and fibre orientation may change the uterine contractions leading
to disturbed peristalsis [46]. The hyperperistalsis induces uterine
auto-traumatisation and desquamation of basal endometrium
which is transported into the peritoneal cavity [7]. Basal
endometrium has an increased potential for implantation and
proliferation resulting in pelvic endometriosis [7]. In addition,
traumatization of the basal endometrium and the JZ could allow
endometrial glands topenetrate into themyometrium anddevelop
adenomyosis [7]. In particular, infiltrating endometriosis might be
related to adenomyosis due to the infiltrating growth pattern.
The JZ is easily visualized by MRI. Abnormal widening (diffuse
or focal) of the innermyometrium or JZ is one of the MRI features
associated with adenomyosis (Fig. 1). It is the consequence of
uncoordinated inner myocyte proliferation called JZ hyperplasia
[5]. TheJZ hyperplasia and accompanying disruption could initiate
endometrial mucosal penetration of endometrial glands into the
myometrium [8].
MRI is highly accurate in the diagnosis of uterine adenomyosis
[911]. At MRI the heterotopic endometrial tissue may be seen as
small foci of increased signal intensity in theJZ. This signhas ahigh
diagnostic specificity for adenomyosisbut cannot stand as the only
criterion, as it may only be seen in less thanhalf the cases. In peri-
and post-menopausal women, a JZ thickness of !12mm was
established as the optimal isolated criterion for adenomyosis [9].
TheJZ thickness, however, is hormone dependent and increases in
the premenopause [12], and therefore other criteri describing the
invasion of the JZ related to unaffected JZ or total uterine wall
thickness should be added [10,11]. With the use of these criteria in
combination, MRI is highly accurate in the diagnosis of uterine
adenomyosis [13].
European Journal of Obstetrics & Gynecology and Reproductive Biology 157 (2011) 206211
A R T I C L E I N F O
Article history:
Received 1 September 2010
Received in revised form 21 December 2010
Accepted 6 March 2011
Keywords:
Adenomyosis
Endometriosis
Magnetic resonance imaging
A B S T R A C T
Objectives: To evaluate image findings in the junctional zone (JZ) in patients with endometriosis and
correlate with image findings of adenomyosis. To attempt a correlation of the degree of adenomyotic
infiltration with the degree of infiltration and stage of endometriosis.
Study design: Magnetic resonance imaging (MRI) of the uterus was performed in 153 women with
suspecteddeeply infiltratingendometriosisandplanned surgery,and in a referencegroup of129women
without endometriosis, verified during hysterectomy. Changes in theJZ and endometriosis in the pelvis
were described in detail. Diagnosis of adenomyosis at MRI was based on optimal criteria derived from
the hysterectomy control group. The stage of endometriosis (AFS stage) wasdetermined during surgery.
Results: In the group ofwomenwith endometriosis 34.6% had adenomyosis compared with19.4% in the
reference group (p
8/3/2019 Ad en Om Yo Sis
2/6
The association between endometriosis and adenomyosis has
beenevaluated inonly a fewstudies [1416] and in only one study
with optimal criteria [14]. In the present study, the occurrence of
adenomyosis and image changes in the JZ was assessed by MRI in
patientswith rectovaginal endometriosis, as compared to findings
in patients with other forms of endometriosis, and in patients
without the disease.
2.
Material and
methods
2.1. Patients
2.1.1. Group 1: patients with suspected rectovaginal endometriosis
(N = 153)
From 1 January 2001 until 1 June 2005, 153 patients were
referred for MRI and subsequent laparoscopy due to suspected
deeply infiltrating endometriosis. In this period all patients with
suspicion of deeply infiltrating endometriosis were routinely
referred for preoperative MRI to map out and describe the relation
of endometriotic nodules to the rectum and ureters. All patients
were booked fordiagnosticlaparoscopy and laparoscopicresection
of all visible endometriotic lesions. Patients with laparoscopically
confirmed
rectovaginal
endometriosis
were
treated
mainly
with
shaving of endometriotic tissue from the bowel wall, and discoid
resection when needed. The operative findings represented the
definitive diagnosis of endometriosis. Deeply infiltrating endome-
triosis was defined as more than 5 mm invasion (assessed during
surgery) of endometriosis into underlying tissues. The AFS stage
and extent of endometriosis were determined during surgery
according to the revised classification of the American Society for
Reproductive Medicine (1996) [17].
2.1.2. Group 2: patients with cervical cancer (N =29)
These women participated in a study concerning urological
complications following radical hysterectomy. The women under-
wentMRIbefore surgery (1December2001until1April2005), and
the occurrence of endometriosis was noted peroperatively.
2.1.3. Group 3: patients undergoing hysterectomy for benign
conditions (N = 100)
This group consisted of all consecutive pre-menopausal women
who had a hysterectomy due to a benign condition at Aarhus
University Hospital, from September 1998 to February 2000. The
study population included 178 patients. Three were not invited to
participate because of language problems, 14 could not be reached
by phone for an appointment, 53 declined the invitation, and 2
patients were excluded because the uterus was morcellated athysterectomy. All patients underwent MRI followed by hysterec-
tomy within 14 days. The prevalence of adenomyosis in the
excluded patients was 22%, which was no different from the
prevalence in the included patients. In six patients, endometriosis
was diagnosed during surgery. These patientswere excluded from
the present study, leaving 100 patients without endometriosis for
analysis. The main indications for surgery were: abnormal
bleeding 51, symptomatic myomas 35, lower abdominal pain 11
(9 of these 11 patients had concomitant myomas or abnormal
bleeding), dysplasia and borderline ovarian tumour 3. MRI
diagnoses of adenomyosis based on different MRI criteria were
compared with the findings of the pathologic examinations. The
experience of our team for evaluation of adenomyosis, with a high
accuracy ofMRI fordiagnosis of adenomyosis,hasbeenestablishedin this previous study [10], and we used MRI criteria with the
histologically confirmed highest diagnostic accuracy of 84%
(sensitivity 65%, specificity 89%).
2.2. MRI
Before 2000, MRI was performed with 1.5 T scanners (Signa,
General Electric Medical systems, Milwaukee, WI and Gyroscan
ACS.NT, Philips). We acquired 4-mm slices with 1-mm spacing in
the sagittal, coronal, and axial planes relative to the orientation of
the uterine cavity, using T2-weighted fast (turbo) spin echo
sequences (TR/TeEf,35004000ms/90ms, echo train length16) in
all three planes using a matrix of 512 448.We used surface coils
(phase
array
pelvic coils) for data
collection and
completed
theexamination in 3045 min. After 2000, MRI was performed with
new 1.5 T scanners (Signa, Twin-Speed, General Electric Medical
systems, and Achieva, Philips). We optimized our sequences in
each system, which gave us different settings of the sequences in
the two systems. The Philips system provided 4 mm slices with
0.5 mm spacing in the sagittal, coronal, and axial planes relative to
theorientationof theuterine cavity, using T2-weighted fast (turbo)
spin echo sequences (TR/TeEf, 35004000ms/110ms, echo train
length 22) in all three planes. We used a surface coil (sense cardiac
phase array) for data collection using a matrix of 512 448. The
GeneralElectricsystemprovided4 mm sliceswith 0.5-mm spacing
in the sagittal, coronal, and axial planes relative to the orientation
of the uterine cavity, using T2-weighted fast (turbo) spin echo
sequences (TR/TeEf,35004000ms/90ms, echo
train length12) in
Fig. 1. Magnetic resonance imaging. (a) Normal uterus, (b) Adenomyosis.
S.B. Larsen et al./ European Journal of Obstetrics & Gynecology and Reproductive Biology 157 (2011) 206211 207
8/3/2019 Ad en Om Yo Sis
3/6
all tree planes. We used a pelvic phased array surface coil for data
collection using a matrix of 512 448. All examinations were
completed in 3045 min.
The thickness was measured at the thinnest (JZ-min) and
thickest (JZ-max) parts of the anterior and posterior wall in the
sagittal slices. The differencebetween JZ-max and JZ-min (JZ-dif)
was calculated for the anterior and posterior border. The largest
parameter, either anterior or posterior, was used in all calcula-
tions. For each patient all areas with poorly defined margins
suspected of being adenomyosis were described. For these areas
we registered their size, the JZ-max, and presence of high
signal foci.
In patientswith endometriosis the maximal anterior (AW) and
posterior (PW) uterine wall thicknesswasmeasured and invasion
depth of the anterior and posteriorwall was calculated as JZ-max/
maximum wall thickness. The largest invasion depth in either the
anterior or posterior wall was used. In the reference group of
women this parameter was inappropriate as several patients had
myomas, which increased AW and PW.
Adenomyosis was thought to be present: (a) in the presence of
focal poorly demarcated low intensity areas in the myometrium
with high intensity myometrial spots arising from the endometrial
myometrial boarder, or (b) with >15 mm junztional thickness, or
(c) when a JZ-dif of >5 mm was present.At MRI the presence and size of infiltrating recto-vaginal
endometriosis were measured in three perpendicular planes (d1,
d2, d3) and the relation to rectum and ureters was described.
Volume of infiltrations was calculated according to ellipse volume
p/6 d1 d2 d3. MRI scans were evaluated by the same MRI
specialist (EL).
2.2.1. Data analysis and statistics
The statistical analyses were performed using X2, Fishers exact
test (F), and KruskalWallis test (KW) whennon-parametrictests
were appropriate. Median and 1090 percentiles (p10p90) were
used for distributions where means and standard deviations (SD)
were unsuitable.MantelHaenszel test wasused when two groups
were compared and adjusted for control variables. The group of
patients with endometriosis (group 1) were compared to the
groups of controls (groups 2 + 3) in the analysis.
3. Results
Mostwomenwithendometriosishad severe infiltrating disease
(Table1). Thewomenwithendometriosiswere younger, had fewer
children and were more often on hormone therapy.
The prevalence of adenomyosis in the group of women with
endometriosis was34.6%, and higher than the prevalence found in
the control group (groups 2 + 3) (19.4%) (Table 2). Among women
with endometriosis,more women had an irregular JZ compared to
the control group. Moreover the irregularity was more pro-
nounced, with higher values of JZ-dif in patients with endometri-
osis.
The JZ was not so broad among endometriosis patients (lower
median ofJZ-max). Fifty percent of the womenwithendometriosis
had a JZ-max of 7 mm or lower. Among women without
adenomyosis, the group of endometriosis patients had a signifi-
cantly thinner JZ-max compared with group (2 + 3) (median, p10
p90: 6.0 mm, 3.010.4, vs 9.0 mm, 5.012.0) (p
8/3/2019 Ad en Om Yo Sis
4/6
Among womenwith severe endometriosis (AFS stage IV) 42.8%
had adenomyosis compared to 29.4% among women in the other 3
stages (AFS stages I + II + III) (p = 0.10) (Table 3). Deeper wall
invasion and JZ-dif were seen in more women with AFS stage IV
compared to stages IIII (Table 4). Adjustment for the presence of
endometriomas did not change the estimates for AFS stage.
MRI revealed deeply infiltrating recto-vaginal endometriosis
among 75.8%of theendometrioticpatients,and 34.5%of thesehadadenomyosis compared to 35.1% in the group without recto-
vaginal endometriosis (p > 0.05). There were no more cases of
adenomyosis in patients with large infiltrations, and the depth of
infiltration of adenomyosis was no deeper in patients with large
volumes of infiltrations (Table 3).No more patientswithboth AFS
stage IVand rectovaginal infiltrations had adenomyosis, and there
was no deeper infiltration of adenomyosis in these patients.
4. Comments
One third of young women with clinically suspected deeply
infiltrating endometriosis had MRI findings of uterine adenomyo-
sis. Symptomatic and severe infiltrating endometriosis seems to becorrelated with adenomyosis and should motivate a diagnostic
evaluation of adenomyosis among these patients. Persistence of
dysmenorrhoea and non-menstrual pain after optimal surgical
resection of peritoneal endometriosis are more likely in patients
with increasing JZ thickness suggesting adenomyosis [1921].
Postoperative treatment of these patients may thus be needed.
Moreover, adenomyosis may be an important cause of infertility
[7,22], which seems to improve after proper treatment [23].
Classic adenomyosis is present in 2035% of patients
undergoing hysterectomy [24], and is more commonly diagnosed
in the forties or fifties, whereas endometriosis is diagnosed in
younger age groups [25]. Younger patientswithendometriosis had
a thin JZ, whereas the control group of older womenhad a broader
Table 2
Characteristics of the junctional zone (JZ) in the three groups of women.
Group 1 Group 2 Group 3 Significance
Patients with
endometriosis (N=153)
Patients with cervical
cancer (N=29)
Patients who had a
hysterectomy (N=100)
(p)**
Adenomyosis
Yes, N (%) 53 (34.6%) 6 (20.7%) 19 (19.0%)
8/3/2019 Ad en Om Yo Sis
5/6
JZ. The JZ increase with age before the menopause [12,26] and a
regular broader symmetric JZ may most likely just be a hormone-
dependent age-related change [27]. It may have clinical signifi-
cance but shouldbe separated from, and seemsnot to be related to,
adenomyosis [28].
Adenomyosis requires infiltration of endometrial glands and
stroma into the myometrium, and image reflection of invasion as
jzmax/wall thickness >40% [11] or JZ-diff >5 [10] should be used.
The latter might not correlate to patients custom JZ and be more
appropriate in the presence of myomas. It should be distinguished
from uterine contractions, which are seen as transient regular
swellings of the JZ. Even these measures, however, should be
evaluated in a young age group with histopathology for verifica-
tion. In younger patients with endometriosis, adenomyosis was
seen as localized irregular burst in a thin JZ. It could indicate that
adenomyosis is initiated by a primary break in the endomyome-
trial border followed by, but not preceded by, localised muscular
JZ-hypertrophy. It may be caused by intrinsic auto-traumatic
factors [7,29], or external traumatisation by, for example,
pregnancy [30].
Nevertheless, although JZ thickness differed, the depth of
invasion of adenomyosis was the same in patients with
adenomyosis in the group with endometriosis compared to the
control group. Thus image findings in this young population ofpatients might most likely just be an earlier manifestation of
adenomyosis found in the older populationwithadenomyosis, and
JZ changes in endometriosis are not histologically verified and may
constitute disease other than adenomyosis [28]. More studies are
needed to clarify the cause of different image findings.
The association between endometriosis and adenomyosis has
beenevaluated inonly a fewstudies.Ourfindingswere in linewith
the finding in another study [14] but differed from the results in a
study of infertility patients [15], where the prevalence of
adenomyosis was 79% and 28% in patients with and without
endometriosis, respectively. This was unexpected since the
majority of our patients had deeply infiltrating disease, where
more aggressive adenomyosis might have been expected. The
diverging findings might be due to different MRI criteria for thediagnosis of adenomyosis, which are still controversial. Kunz et al.
used aJZ of 10mm fordiagnosisof adenomyosis [15] in contrast to
others, where MRI findings were correlated with histopathology
[9,10,31]. Our use of a restrictive MRI diagnosis of adenomyosis
compared to the criteria proposed by Reinhold et al. [32] resulted
in a lower prevalence of adenomyosis in both groups without
changing the difference between the groups.
Adenomyotic changes were not evident in two thirds of the
patients with endometriosis, and the presence and size of
rectovaginal infiltrating endometriosis was not correlated with
adenomyosis or depth of infiltration of adenomyosis. The theory of
endometriosis as a primary disease of the archiometra [7] was not
clearly reinforced in this study, as no correspondence in level of
invasive
potential
in the
myometrium
and
peritoneum
was seen.This goes against a common intrinsic abnormality in eutopic and
ectopic endometrium. There could, however, be different expres-
sionsof invasive potential dependent on local factorsaccounting for
the different findings.
Nevertheless, in line with Kunz et al. [15], adenomyosis seemed
tobemore invasive inAFSstage4.TheAFS scoredoesnotaddress the
clinically most important extentofdisease whichis deep infiltrating
endometriosis. The AFS score corresponds more with the inflam-
matory and adhesive components of endometriosis and with
endometriomas. Dysperistalsis and menorrhagia in adenomyosis
couldgive rise toa larger loadofperitoneal endometrial cells during
menstruation,which couldpromoteadhesionand inflammationand
account for this finding, but this inflammation did not seem to give
rise
to
more
deep
infiltration.
The optimal control group would have been an age-matched
group of patients with no clinical symptoms. It is very difficult and
expensive, however, to establish such a group with a sufficient
number of patients, and no histopathology verification can be
established. The diagnostic criteria at MRI for diagnosis of
adenomyosis are still controversial, and motivate our use of a
control group with histopathology confirmation of the diagnostic
criteria used [13]. Thus the prevalence of adenomyosis would be
expected to be lower in a control group of younger asymptomatic
patients compared to theusedcontrol groupofolderpremenopausal
womenundergoinghysterectomy forbenign conditions. Thoughno
endometriosis was seen at histopathology and described during
operation, exclusion of minorendometriosiswould have required a
uniform staging of a single experienced observer.
In spite of the above-mentioned conditions the group ofwomen
with severe endometriosis demonstrated a higher prevalence of
adenomyosis than the control groups and illustrates the need for
an imaging technique for diagnosis of adenomyosis in patients
with endometriosis. This should be by MRI or transvaginal
ultrasound (TVS) by a clinician skilled in the sonographic findings
of adenomyosis. The diagnostic accuracy of TVS [33] is in line with
MRI [13]. TVS is very observer-dependent in the evaluation of
adenomyosis [34]. MRI has the advantage of being able to predict
deep infiltrating endometriosis at all locations even outside thepelvis and to define the exact extent of both endometriosis and
adenomyosis [35,36].
In summary,in this study a systematic description ofJZ changes
in endometriosis implied an association of severe symptomatic
endometriosis with adenomyosis, but the invasive potential of
endometrial cells in the uterusand peritoneum corresponded only
to a limited extent.
References
[1] Hudelist G, Keckstein J,WrightJT. The migrating adenomyoma: past views onthe etiology of adenomyosis andendometriosis. FertilSteril2009;92:153643.
[2] Benagiano G, Brosens I. History of adenomyosis. Best Pract Res Clin ObstetGynaecol 2006;20:44963.
[3] Leyendecker G, Kunz
G, Noe
M, Herbertz M, Mall
G. Endometriosis: adysfunction and disease of the archimetra. Hum Reprod Update1998;4:75262.
[4] UduwelaAS,Perera MA,Aiqing L, Fraser IS. Endometrial-myometrial interface:relationship to adenomyosis and changes in pregnancy. Obstet Gynecol Surv2000;55:390400.
[5] Brosens JJ, de SN, Barker FG. Uterine junctional zone: function and disease.Lancet 1995;346:55860.
[6] LeyendeckerG,KunzG,Kissler S,Wildt L. Adenomyosis and reproduction.BestPract Res Clin Obstet Gynaecol 2006;20:52346.
[7] Leyendecker G, Kunz G, Herbertz M, et al. Uterine peristaltic activity and thedevelopment of endometriosis. Ann N Y Acad Sci 2004;1034:33855.
[8] Brosens JJ, Barker FG, de SN. Myometrial zonal differentiation and uterinejunctional zone hyperplasia in the non-pregnant uterus. Hum Reprod Update1998;4:496502.
[9] Reinhold C, McCarthy S, Bret PM, et al. Diffuse adenomyosis: comparison ofendovaginal US and MR imaging with histopathologic correlation. Radiology1996;199:1518.
[10] Dueholm M, Lundorf E, Hansen ES, Ledertoug S, Srensen JS, Olesen F.
Magnetic resonance
imaging
and
transvaginal
ultrasonography
for diagnosisof adenomyosis. Fertil Steril 2001;76:58894.
[11] Bazot M, Cortez A, Darai E, et al. Ultrasonography compared with magneticresonance imaging for the diagnosis of adenomyosis: correlation with histo-pathology. Hum Reprod 2001;16:242733.
[12] Hauth EA, Jaeger HJ, Libera H, Lange S, Forsting M.MR imaging of the uterusand cervix in healthy women: determination of normal values. Eur Radiol2007;17:73442.
[13] Dueholm M, Lundorf E. Transvaginal ultrasound or MRI for diagnosis ofadenomyosis. Curr Opin Obstet Gynecol 2007;19:50512.
[14] Bazot M, Fiori O, Darai E. Adenomyosis in endometriosisprevalence andimpact on fertility. Evidence from magnetic resonance imaging. Hum Reprod2006;21:11012.
[15] Kunz G, Beil D, Huppert P, Noe M, Kissler S, Leyendecker G. Adenomyosis inendometriosisprevalence and impact on fertility. Evidence from magneticresonance imaging. Hum Reprod 2005;20:230916.
[16] Zacharia TT, ONeill MJ. Prevalence and distribution of adnexal findingssuggesting endometriosis in patients with MR diagnosis of adenomyosis. BrJ Radiol 2006;79:3037.
S.B. Larsen et al./ European Journal of Obstetrics & Gynecology and Reproductive Biology 157 (2011) 206211210
8/3/2019 Ad en Om Yo Sis
6/6
[17] Revised American Society for Reproductive Medicine classification of endo-metriosis: 1996. Fertil Steril 1997;67:81721.
[18] Reinhold C, Tafazoli F, Mehio A, et al. Uterine adenomyosis: endovaginal USand MR imaging features with histopathologic correlation. Radiographics1999;19 Spec No: S14760.
[19] Landi S, Mereu L, Pontrelli G, et al. The influence of adenomyosis in patientslaparoscopically treated for deep endometriosis. J Minim Invasive Gynecol2008;15:56670.
[20] Parker JD, Leondires M, Sinaii N, Premkumar A, Nieman LK, Stratton P.Persistence of dysmenorrhea and nonmenstrualpain afteroptimal endome-triosis surgery may indicate adenomyosis. Fertil Steril 2006;86:7115.
[21] Ferrero
S, Camerini
G, Menada
MV, Biscaldi
E, Ragni
N, Remorgida V.Uterine adenomyosis in persistence of dysmenorrhea after surgical exci-sion of pelvic endometriosis and colorectal resection. J Reprod Med2009;54:36672.
[22] KimMD, Kim S,Kim NK, et al. Long-term results ofuterine artery embolizationfor symptomatic adenomyosis. AJR Am J Roentgenol 2007;188:17681.
[23] Mijatovic V, Florijn E, Halim N, Schats R, Hompes P. Adenomyosis has noadverse effects on IVF/ICSI outcomes in women with endometriosis treatedwith long-term pituitary down-regulation before IVF/ICSI. Eur J Obstet Gyne-col Reprod Biol 2010;19.
[24] Vercellini P, Vigano P, Somigliana E, Daguati R, Abbiati A, Fedele L. Adeno-myosis: epidemiological factors. Best Pract Res Clin Obstet Gynaecol2006;20:46577.
[25] Templeman C, Marshall SF, Ursin G, et al. Adenomyosis and endometriosis inthe California Teachers Study. Fertil Steril 2008;90:41524.
[26] Kunz G, Herbertz M, Beil D, Huppert P, Leyendecker G. Adenomyosis as adisorder of the early and late human reproductive period. Reprod BiomedOnline 2007;15:6815.
[27] Tamai K, Togashi K, Ito T, Morisawa N, Fujiwara T, Koyama T. MR imagingfindings of adenomyosis: correlation with histopathologic features and diag-nostic pitfalls. Radiographics 2005;25:2140.
[28] Tocci A, Greco E, Ubaldi FM. Adenomyosis and endometrialsubendometrialmyometrium unit disruption disease are two different entities. ReprodBiomed Online 2008;17:28191.
[29] Giudice LC, Kao LC. Endometriosis. Lancet 2004;364:178999.[30] Vercellini P, Parazzini F, Oldani S, Panazza S, Bramante T, Crosignani PG.
Adenomyosis at hysterectomy: a study on frequency distribution and patientcharacteristics. Hum Reprod 1995;10:11602.
[31] Ascher SM, Arnold LL, Patt RH, et al. Adenomyosis: prospective comparison of
MR imaging
and
transvaginal
sonography.
Radiology
1994;190:8036.[32] Reinhold C, Tafazoli F,Wang L. Imaging featuresof adenomyosis. Hum ReprodUpdate 1998;4:33749.
[33] Meredith SM, Sanchez-Ramos L, Kaunitz AM. Diagnostic accuracy of trans-vaginal sonography for the diagnosis of adenomyosis: systematic review andmetaanalysis. Am J Obstet Gynecol 2009;201:10716.
[34] Dueholm M, Lundorf E, Sorensen JS, Ledertoug S, Olesen F, Laursen H. Repro-ducibility of evaluation of the uterus by transvaginal sonography, hysteroso-nographic examination,hysteroscopy and magnetic resonance imaging. HumReprod 2002;17:195200.
[35] Bazot M, Bornier C, Dubernard G, Roseau G, Cortez A, Darai E. Accuracy ofmagnetic resonance imaging and rectal endoscopic sonography for the pre-diction of location of deep pelvic endometriosis. Hum Reprod 2007;22:145763.
[36] Bazot M, Detchev R, Cortez A, Amouyal P, Uzan S, Darai E. Transvaginalsonography and rectal endoscopic sonography for the assessment ofpelvic endometriosis: a preliminary comparison. Hum Reprod 2003;18:168692.
S.B. Larsen et al./ European Journal of Obstetrics & Gynecology and Reproductive Biology 157 (2011) 206211 211