Diagnosis of Endometriosis

8.02.2019
1
Diagnosis of Endometriosis
Current Techniques and New Technologies
TEVFİK YOLDEMİR MD. BSc. MA. PhD.
tyoldemir
profdrdryoldemir
Primary locations of endometriosis, their
prevalence in patients with endometriosis
Milwaukee, WI, USA), by using an eight-channel high-reso-
lution surface phased-array torso coil with array spatial sensi-
tivity technique (ASSET) parallel acquisition.
In preparation for imaging, it is recommended that patients
fast (4–6 h) before the examination. Bowel preparation in-
cludes an enema administered approximately 2–3 h before
the examination. The study should not be conducted during
the menstrual cycle.
MR imaging is performed with moderate repletion of the
patient’s bladder, since an overfilled bladder may cause
detrusor contractions and may obliterate the adjacent recesses
thus compromising the identification of small parietal nodules
[1, 2]. On the other hand, an empty bladder prevents optimal
visualization of the ureters.
MR imaging is performed with the patient lying in the
supine position (entry position feet first). In patients who show
a dilatation of the excretory system, the urographic phase is
acquired in the prone position. In claustrophobic patients,
prone position may reduce anxiety and improve exam
acceptability.
When the clinical evaluation suggests a rectosigmoid en-
dometriosis, rectal opacification is performed before the ex-
amination. Retrograde distension of the rectum and the sig-
moid colon is obtained inside the gantry with a rectal enema
of 750 mL of saline solution introduced through a Nelaton
catheter (20 Ch, 6.67 mm × 360 mm). Bowel cleansing is
performed through oral administration of a polyethylene gly-
col solution (1000 mL) the day before the study. In these
patients the intravenous administration of an antispasmodic
agent, scopolamine-N-butyl bromide (Buscopan® 20 mg;
Boehringer Ingelheim, Milano, Italy) just before image ac-
quisition is mandatory to reduce motion artefacts caused by
bowel peristalsis. Even if rectal opacification is not strictly
necessary to detect endometriotic lesions of the intestinal
wall, rectal distension may be useful to evaluate the degree
of bowel stenosis.
Table 1 Primary locations of endometriosis, their prevalence in patients with endometriosis, clinical features and differential diagnosis [3, 10, 11]
Locations % Clinical features MR differential diagnosis
Bladder 6.4–20 dysuria, hematuria, urinary storage symptoms,
suprapubic pain
urachal remnant, epithelial and mesenchymal
tumours
Ureters 0.01–1 dysmenorrhea, dyspareunia, flank pain
(hydronephrosis)
obstruction by cervical cancer
Ovaries 20–40 nonspecific pelvic pain dermoids, hemorrhagic cysts, endometrioid
and clear cell tumours
Round ligaments 0.3–14 painful inguinal mass, nonspecific pelvic pain
Retrocervical region,
uterosacral ligaments
69.2 painful symptoms, dyspareunia peritoneal metastases
Vagina 14.5 dysmenorrhea, dyspareunia, postcoital spotting cervical and vaginal carcinoma
Rectosigmoid colon 9.9–37 dyschezia, cyclic pain, rectal bleeding colorectal cancer, metastatic implants
Insights into Imaging (2018) 9:149–172
Fourbasic sonographic steps forexaminingwomen with
clinical suspicion of deep infiltratingendometriosis (DIE)
320 Guerriero and Condous et al.
the presence or absence of vaginal and/or low rectal
endometriosis7
. The pelvic examination should include
speculum examination (direct visualization of vaginal or
cervical DIE) and vaginal palpation. Mobility, fixation
and/or tenderness of the uterus should be evaluated
carefully. Site-specific tenderness in the pelvis should also
be evaluated.
SONOGRAPHY OVERVIEW
The purpose of performing an ultrasound examination
in a woman with suspected endometriosis is to try to
explain underlying symptoms, map the disease location
and assess the severity of disease prior to medical
therapy or surgical intervention. Various ultrasound
approaches have been published, but to date none has
been externally validated16,17
. We propose four basic
sonographic steps when examining women with suspected
or known endometriosis, as shown in Figure 1. Note that
these steps can be adopted in this or any order as long
as ALL four steps are performed to confirm/exclude the
different forms of endometriosis.
Using TVS as the first-line imaging tool, the operator
should examine the uterus and the adnexa. The mobility
of the uterus should be evaluated: normal, reduced
or fixed (‘question mark sign’)18
. Sonographic signs
of adenomyosis should be searched for and described
using the terms and definitions published in the
Morphological Uterus Sonographic Assessment consensus
opinion19
.
The presence or absence of endometriomas (Figure S1a),
their size, measured systematically in three orthogonal
planes (see ‘Measurement of lesions’, below), the number
of endometriomas and their ultrasound appearance
should be noted20. The sonographic characteristics of any
endometrioma should be described using the International
Ovarian Tumor Analysis terminology21
. An atypical
endometrioma (Figure S1b) is defined as a unilocular-solid
First step
Routine evaluation of uterus and adnexa
(+ sonographic signs of adenomyosis/presence or
absence of endometrioma)
Evaluation of transvaginal sonographic
‘soft markers’
(i.e. site-specific tenderness and ovarian mobility)
Assessment of status of POD using real-time
ultrasound-based ‘sliding sign’
Assessment for DIE nodules in
anterior and posterior compartments
Second step
Third step
Fourth step
Dynamicultrasonography
Figure 1 Four basic sonographic steps for examining women with
clinical suspicion of deep infiltrating endometriosis (DIE) or known
endometriosis. All steps should be performed, but not necessarily in
this order. Note, bladder should contain small amount of urine.
Dynamic ultrasonography is when the operator performing the
ultrasound examination assesses both the pelvic organs and their
mobility in real-time. POD, pouch of Douglas.
mass with ground glass echogenicity with a papillary
projection, a color score of 1 or 2 and no flow inside the
papillary projection20
.
Ovarian endometriomas are associated frequently
with other endometriotic lesions, such as adhesions
and DIE22,23. The ‘kissing’ ovaries sign (Figure S2)
suggests that there are severe pelvic adhesions; bowel
and Fallopian tube endometriosis are significantly more
frequent in women with kissing ovaries vs those without
kissing ovaries: 18.5% vs 2.5% and 92.6% vs 33%,
respectively24
.
Endometriomas may undergo decidualization in preg-
nancy, in which case they can be confused with an ovarian
malignancy on ultrasound examination (Figure S3)25.
Simultaneous presence of other endometriotic lesions
may facilitate a correct diagnosis of endometrioma in
pregnancy and minimize the risk of unnecessary surgery.
The second step is to search for sonographic ‘soft
markers’, i.e. site-specific tenderness (SST) and fixed
ovaries. The presence of soft markers increases the likeli-
hood of superficial endometriosis and adhesions26,27
. By
applying pressure between the uterus and ovary, one can
assess if the ovary is fixed to the uterus medially, to the
pelvic side wall laterally or to the USLs. The presence of
adhesions can also be suspected if, on palpation with the
probe and/or abdominal palpation with the free hand,
the ovaries or the uterus appear to be fixed to adjacent
structures (broad ligament, POD, bladder, rectum and/or
parietal peritoneum). If there is pelvic fluid, fine strands
of tissue (adhesions) may be seen between the ovary
(with or without endometrioma) and the uterus or the
peritoneum of the POD27–30
.
If there are endometriomas or pelvic endometriosis, the
Fallopian tubes are frequently involved in the disease
process. Adhesions may distort the normal Fallopian
tubal course and occlusion of the Fallopian tube(s) by
endometriotic foci or distal tubular adhesions may also
occur. As a consequence, a sactosalpinx may develop. For
these reasons, hydrosalpinx/hematosalpinx and peritoneal
cysts should be searched for and reported.
The third step is to assess the status of the POD
using the real-time TVS-based ‘sliding sign’. In order
to assess the sliding sign when the uterus is anteverted
(Figure 2a), gentle pressure is placed against the cervix
using the transvaginal probe, to establish whether the
anterior rectum glides freely across the posterior aspect
of the cervix (retrocervical region) and posterior vaginal
wall. If the anterior rectal wall does so, the ‘sliding sign’ is
considered positive for this location (Videoclip S1a). The
examiner then places one hand over the woman’s lower
anterior abdominal wall in order to ballot the uterus
between the palpating hand and the transvaginal probe
(which is held in the other hand), to assess whether the
anterior bowel glides freely over the posterior aspect of the
upper uterus/fundus. If it does so, the sliding sign is also
considered positive in this region (Videoclip S1b). When
the sliding sign is found to be positive in both of these
anatomical regions (retrocervix and posterior uterine
fundus), the POD is recorded as being not obliterated.
Copyright © 2016 ISUOG. Published by John Wiley & Sons Ltd. Ultrasound Obstet Gynecol 2016; 48: 318–332.
Ultrasound Obstet Gynecol 2016; 48: 318–332
Agreement between Observers A and B
Observer agreement on TVS 739
Table 1 Agreement between Observers A and B, and between each observer and laparoscopy, for the diagnosis by transvaginal sonography
(TVS) of deep infiltrating endometriosis (DIE) in different pelvic locations
Gwet’s AC1 (95% CI)
Agreement between
DIE location
Prevalence*
(n (%))
Observers A and B
on TVS
Observer A on TVS
and laparoscopy
Observer B on TVS
and laparoscopy
Observers A and B
on TVS (%)
Vagina 11 (17) 0.93 (0.86–1.00) 0.82 (0.70–0.95) 0.89 (0.77–0.99) 95.4
Bladder 4 (6) 1.00 (1.00–1.00) 0.95 (0.89–1.00) 0.95 (0.89–1.00) 100
USL 17 (26) 0.84 (0.72–0.97) 0.72 (0.56–0.89) 0.69 (0.51–0.86) 89.2
Adnexa 18 (28) 0.95 (0.87–1.00) 0.76 (0.60–0.92) 0.76 (0.60–0.92) 97.0
RVS 8 (12) 0.95 (0.89–1.00) 0.83 (0.71–0.95) 0.82 (0.70–0.94) 95.4
Rectosigmoid 16 (25) 0.98 (0.93–1.00) 0.93 (0.84–1.00) 0.95 (0.88–1.00) 98.5
*Prevalence determined by histopathological findings at laparoscopy. RVS, rectovaginal septum; USL, uterosacral ligament.
Table 2 Diagnostic performance and accuracy of diagnosis by transvaginal sonography of deep infiltrating endometriosis (DIE) by
Observers A and B, with respect to findings on laparoscopy
DIE location Observer
Sensitivity
(% (95% CI))
Specificity
(% (95% CI))
PPV
(% (95% CI))
NPV
(% (95% CI))
LR+
(95% CI)
LR–
(95% CI)
Accuracy
(%)
Vagina A 62 94 73 91 10.7 0.4 88
(35–88) (88–100) (46–99) (83–98) (3.3–35.0) (0.20–0.81)
B 82 94 75 96 14.7 0.2 92
(56–100) (88–100) (51–100) (91–100) (4.7–45.8) (0.05–0.68)
Bladder A 67 97 50 98 20.7 0.34 95
(13–100) (92–100) (1.0–99) (95–100) (4.3–100.2) (0.01–1.7)
B 67 97 50 98 20.7 0.34 95
(13–100) (92–100) (1.0–99) (95–100) (4.3–100.2) (0.01–1.7)
USL A 73 83 47 94 4.4 0.3 82
(46–99) (73–93) (23–71) (87–100) (2.2–8.8) (0.12–0.89)
B 53 90 64 84 5.1 0.5 80
(29–77) (81–98) (39–89) (74–94) (2.0–13.1) (0.31–0.88)
Adnexa A 71 93 83 87 10.5 0.3 86
(52–91) (86–100) (66–100) (78–97) (3.4–32.3) (0.16–0.61)
B 71 93 83 87 10.5 0.31 86
(52–91) (86–100) (66–100) (78–97) (3.4–32.3) (0.16–0.61)
RVS A 40 90 25 95 4.0 0.7 86
(3.0–83) (82–98) (3.2–65) (89–100) (1.1–14.9) (0.30–1.4)
B 33 87 11 96 2.6 0.8 85
(20–87) (79–95) (9–32) (92–100) (0.5–14.5) (0.34–1.71)
Rectosigmoid A 93 96 88 98 23.3 0.1 95
(81–100) (91–100) (71–100) (94–100) (6.0–91.3) (0.01–0.5)
B 94 98 94 98 45.9 0.1 97
(82–100) (94–100) (82–100) (94–100) (6.5–321.0) (0.01–0.43)
LR+, positive likelihood ratio; LR–, negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive value; RVS,
rectovaginal septum; USL, uterosacral ligament.
extent of DIE is questionable. Egekvist et al. evaluated
the interobserver variation of TVS measurements of the
size of DIE lesions in the rectosigmoid wall in 24 women
with rectal DIE20
. Examinations were performed by one
experienced and one less experienced sonographer. High
agreement was observed concerning the identification
of rectosigmoidal endometriosis. The probability of
differences in size within 30% of the mean was 0.81,
0.63 and 0.61 for length, width and depth, respectively.
In our study, we aimed to validate previous results
in a large number of patients, and to determine
interobserver agreement for examiners with statistically
proven high-level experience in TVS for DIE, as
demonstrated previously10. In the present study, Gwet’s
AC1 levels for diagnosis of affected anatomical sites
using TVS demonstrate almost perfect (0.81–1.00)
interobserver agreement. In addition, we were able
to underline the high diagnostic accuracy of TVS
for diagnosing DIE when compared to laparoscopy,
comparable to evidence published previously9. The high
interobserver agreement demonstrated in this study
may be explained by the proven high skill level of
both observers. A recent publication by our group
demonstrated the necessity of at least 40 supervised
scans performed in a tertiary referral center setting in
order to be able to diagnose DIE by TVS with high
accuracy10
. For DIE locations that involve urologists
or colorectal surgeons for further surgical treatment,
we were able to show an almost perfect interobserver
agreement in diagnosing DIE of the bladder and the
rectosigmoid. Nevertheless, the results of this study must
be interpreted with caution. Although we included a large
Diagnostic performanceand accuracyof diagnosis by
transvaginal sonographyofdeep infiltratingendometriosis
Observer agreement on TVS 739
Table 1 Agreement between Observers A and B, and between each observer and laparoscopy, for the diagnosis by transvaginal sonography
(TVS) of deep infiltrating endometriosis (DIE) in different pelvic locations
Gwet’s AC1 (95% CI)
Agreement between
DIE location
Prevalence*
(n (%))
Observers A and B
on TVS
Observer A on TVS
and laparoscopy
Observer B on TVS
and laparoscopy
Observers A and B
on TVS (%)
Vagina 11 (17) 0.93 (0.86–1.00) 0.82 (0.70–0.95) 0.89 (0.77–0.99) 95.4
Bladder 4 (6) 1.00 (1.00–1.00) 0.95 (0.89–1.00) 0.95 (0.89–1.00) 100
USL 17 (26) 0.84 (0.72–0.97) 0.72 (0.56–0.89) 0.69 (0.51–0.86) 89.2
Adnexa 18 (28) 0.95 (0.87–1.00) 0.76 (0.60–0.92) 0.76 (0.60–0.92) 97.0
RVS 8 (12) 0.95 (0.89–1.00) 0.83 (0.71–0.95) 0.82 (0.70–0.94) 95.4
Rectosigmoid 16 (25) 0.98 (0.93–1.00) 0.93 (0.84–1.00) 0.95 (0.88–1.00) 98.5
*Prevalence determined by histopathological findings at laparoscopy. RVS, rectovaginal septum; USL, uterosacral ligament.
Table 2 Diagnostic performance and accuracy of diagnosis by transvaginal sonography of deep infiltrating endometriosis (DIE) by
Observers A and B, with respect to findings on laparoscopy
DIE location Observer
Sensitivity
(% (95% CI))
Specificity
(% (95% CI))
PPV
(% (95% CI))
NPV
(% (95% CI))
LR+
(95% CI)
LR–
(95% CI)
Accuracy
(%)
Vagina A 62 94 73 91 10.7 0.4 88
(35–88) (88–100) (46–99) (83–98) (3.3–35.0) (0.20–0.81)
B 82 94 75 96 14.7 0.2 92
(56–100) (88–100) (51–100) (91–100) (4.7–45.8) (0.05–0.68)
Bladder A 67 97 50 98 20.7 0.34 95
(13–100) (92–100) (1.0–99) (95–100) (4.3–100.2) (0.01–1.7)
B 67 97 50 98 20.7 0.34 95
(13–100) (92–100) (1.0–99) (95–100) (4.3–100.2) (0.01–1.7)
USL A 73 83 47 94 4.4 0.3 82
(46–99) (73–93) (23–71) (87–100) (2.2–8.8) (0.12–0.89)
B 53 90 64 84 5.1 0.5 80
(29–77) (81–98) (39–89) (74–94) (2.0–13.1) (0.31–0.88)
Adnexa A 71 93 83 87 10.5 0.3 86
(52–91) (86–100) (66–100) (78–97) (3.4–32.3) (0.16–0.61)
B 71 93 83 87 10.5 0.31 86
(52–91) (86–100) (66–100) (78–97) (3.4–32.3) (0.16–0.61)
RVS A 40 90 25 95 4.0 0.7 86
(3.0–83) (82–98) (3.2–65) (89–100) (1.1–14.9) (0.30–1.4)
B 33 87 11 96 2.6 0.8 85
(20–87) (79–95) (9–32) (92–100) (0.5–14.5) (0.34–1.71)
Rectosigmoid A 93 96 88 98 23.3 0.1 95
(81–100) (91–100) (71–100) (94–100) (6.0–91.3) (0.01–0.5)
B 94 98 94 98 45.9 0.1 97
(82–100) (94–100) (82–100) (94–100) (6.5–321.0) (0.01–0.43)
LR+, positive likelihood ratio; LR–, negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive value; RVS,
rectovaginal septum; USL, uterosacral ligament.
extent of DIE is questionable. Egekvist et al. evaluated
the interobserver variation of TVS measurements of the
size of DIE lesions in the rectosigmoid wall in 24 women
with rectal DIE20
. Examinations were performed by one
experienced and one less experienced sonographer. High
agreement was observed concerning the identification
of rectosigmoidal endometriosis. The probability of
differences in size within 30% of the mean was 0.81,
0.63 and 0.61 for length, width and depth, respectively.
In our study, we aimed to validate previous results
in a large number of patients, and to determine
interobserver agreement for examiners with statistically
proven high-level experience in TVS for DIE, as
demonstrated previously10. In the present study, Gwet’s
AC1 levels for diagnosis of affected anatomical sites
using TVS demonstrate almost perfect (0.81–1.00)
interobserver agreement. In addition, we were able
to underline the high diagnostic accuracy of TVS
for diagnosing DIE when compared to laparoscopy,
comparable to evidence published previously9. The high
interobserver agreement demonstrated in this study
may be explained by the proven high skill level of
both observers. A recent publication by our group
demonstrated the necessity of at least 40 supervised
scans performed in a tertiary referral center setting in
order to be able to diagnose DIE by TVS with high
accuracy10
. For DIE locations that involve urologists
or colorectal surgeons for further surgical treatment,
we were able to show an almost perfect interobserver
agreement in diagnosing DIE of the bladder and the
rectosigmoid. Nevertheless, the results of this study must
be interpreted with caution. Although we included a large
US / Doppler
Almost 50% of the endometriomas had other ultrasound characteristics than the typical ‘unilocular
cyst with ground glass echogenicity of the cyst fluid’.
The rule: ‘premenopausal status, ground glass echogenicity, one to four locules and no papillations
with detectable blood flow’ characterise endometriomas reasonably well, but not as well as subjective
impression [19]. Serum CA-125 levels are not generally useful in distinguishing endometriomas from
other benign tumours and malignancies [19]. They could possibly help to distinguish endometriomas
from other benign lesions [13].
In addition, ultrasound appearance of endometriomas differ between pre- and post-menopausal
women. Endometriomas in the postmenopausal women are less frequently unilocular cysts, and
are less likely to exhibit ground glass echogenicity [19]. Instead, they are more often multilocular-
solid tumours, and more frequently exhibit anechoic cyst fluid or cyst fluid with mixed
echogenicity.
Endometriomas may be misinterpreted, potentially because of the complex echotexture, thick
walls, and solid echogenic appearance of haemorrhagic clots within the endometrioid cystic cavity,
which mimic different dermoid cyst patterns or malignancy. Also, some hyperechoic solid masses (e.g.
fibroids and fibrothecomas) can be misdiagnosed as endometriotic cysts. Inversely, some endome-
triomas can be mistaken for serous cysts, dermoids, and suspected ovarian malignant tumours. The use
of colour or power Doppler detecting the presence, number, and distribution of vessels in the solid
echogenic protrusions of the cyst wall seems to be useful in differentiating endometriomas from
malignant lesions [10,16,17]. Unfortunately the use of colour Doppler evaluation has some limitations ,
because it requires optimal colour Doppler settings, a high quality of the colour Doppler function of the
ultrasound equipment used and an experienced ultrasound examiner. Therefore, the absence of vas-
cularisation does not guarantee that the cyst is benign.
The clinical rule that considers endometrioma a cyst with ground glass echogenicity 1-4 septa, with
no solid parts, in a premenopausal woman, is useful in most clinical settings because it allows the
examiner to skip the colour Doppler assessment of the mass. This clinical rule has almost as good
discriminatory power as the statistically optimal rule, but its sensitivity is lower [19]. Finally all the
principal rules have a sensitivity ranging from 62–73%, a specificity of 94–98%, and a positive predictive
value of 76–89% [19]. Subjective impression by an experienced sonologist was better for identifying
endometriomas (positive predictive value 88.5%, positive likelihood ratio 30.2, sensitivity 81%, speci-
ficity 97%) [10,19]. This is probably because the ultrasound examiner uses other available clinical in-
formation (e.g. pain and dysmenorrhoea) when suggesting a diagnosis, taking into account other
Fig. 2. Atypical ultrasound appearance of an ovarian endometrioma: a unilocular cyst with ground glass echogenicity, internal
papillation and no vascularisation in the papillary projection. This is not a true papillations but hyperechoic tissue consisting of
blood clots or fibrin lying adjacent to the cyst wall.
C. Exacoustos et al. / Best Practice & Research Clinical Obstetrics and Gynaecology 28 (2014) 655–681658
Best Practice & Research Clinical Obstetrics and Gynaecology 28 (2014) 655–681
a unilocular cyst with ground glass echogenicity,
internal papillation and no vascularisation in the
papillary projection.
This is not a true papillations but hyperechoic tissue
consisting of blood clots or fibrin lying adjacent to
the cyst wall.

8.02.2019
2
US / Doppler
Endometriomas and malignancy
Subjective impression can lead to the misclassification of malignancies as endometriomas in 0.2–
0.9% of cases [10,11,19]. Ultrasound characteristics of endometriomas differ in pre- and postmenopausal
women. Masses in postmenopausal women, whose cystic contents have a ground glass appearance,
have a high risk of malignancy.
Endometriomas could serve as precursors of endometrioid borderline ovarian tumours. Endome-
trioid borderline ovarian tumours have the potential to progress to low-grade invasive carcinoma.
Although clear-cell borderline ovarian tumours have been associated with endometriosis, a stepwise
molecular pathway for the progression of endometriosis to clear carcinoma has not yet been identified
[20].
Borderline tumours and carcinomas arising from endometrioid cysts show a vascularised solid
component at ultrasound examination (Fig. 3). The presence of typical sonographic features for ovarian
malignant lesions suggests that benign endometrioid cysts and malignant and borderline tumours
arising from endometriosis might be easier to assess by an expert ultrasound examiner compared with
normal ovarian masses [9]. This is not the case with ovarian cysts found during pregnancy, were dif-
ferentiation between borderline tumours and decidualised endometriotic cysts can be more difficult.
Ultrasound examiners should always take into account the phenomenon of decidualisation, which is a
major contributory factor to incorrect diagnosis in pregnant women [21]. Most decidualised endo-
metriomas (82%) were described as manifesting vascularised rounded papillary projections with a
smooth contour in an ovarian cyst with one or more cyst locules and ground glass, or low level
echogenicity of the cyst fluid [22].
Fig. 3. Three-dimensional ultrasound and power-Doppler image of an endometrioid borderline tumour. Note the irregular papil-
lations and the vascularisation in the papillary projections. These are true papillations composed of borderline malignant solid
tissue.
C. Exacoustos et al. / Best Practice & Research Clinical Obstetrics and Gynaecology 28 (2014) 655–681 659
Three-dimensional ultrasound and
power-Doppler image of an
endometrioid borderline tumour.
Note the irregular papillations and the
vascularisation in the papillary
projections.
These are true papillations composed of
borderline malignant solid tissue.
frequency of different intestinal deep endometriosis
lesions according to Piketty’s et al. study282 A.Massein et al.
clinicopathological types of endometriosis can be distin-
guished, generally intricately linked: superficial periton-
eal endometriosis, ovarian endometriosis (cystic lesions
known as endometriomas) and deep pelvic endometriosis.
Extrapelvic locations (abdominal, pleural) are rare.
Some essential points concerning deep
pelvic endometriosis
There are two non-consensual definitions of deep pelvic
endometriosis: sub-peritoneal implants penetrating for
more than 5 mm under the peritoneum or infiltration of the
uterosacral ligaments and/or the muscles of adjacent pelvic
organs.
Deep pelvic endometriosis is responsible for chronic
pelvic pain (dysmenorrhea, deep dyspareunia), with or with-
out menstrual aggravation, but also dyschezia, rectorrhagia
and dysuria.It is also responsible for infertility with a fre-
quency estimated at 20 to 40% in cases of infertility.Rarely,
endometriosis is asymptomatic.
Its pathogenesis has not been elucidated. Several
hypotheses have been discussed including reflux of
endometrial fragments through the fallopian tubes dur-
ing menstruation, metaplasia of tissues derived from
the coelomic epithelium into endometrial tissue, vascular
and/or lymphatic emboli, and involvement of epigenetic and
environmental factors.
The diagnosis is often made too late, after an average
7 years of pain developing or after several years of medi-
cally assisted procreation treatment.Endometriosis can be
suspected clinically but imaging is the procedure that con-
firms it and maps the lesions. Laparoscopy should not be
performed for diagnostic purposes because it is not without
risk and can ignore some deep lesions or those masked by
adhesions.
Sites of deep endometriosis
Lesions of the uterosacral ligaments are the most common
(Fig.1).They are associated with involvement of the torus
uterinus, which is a small transverse thickening on the pos-
terior surface of the cervix, between the insertions of the
two uterosacral ligaments.
Intestinal and urinary tract locations are the most severe
forms of deep endometriosis (Fig.2).The multiple intestinal
Figure 1. Table showing the frequency of deep pelvic endometriosis according to Chapron’s surgical serie [3].
Figure 2. Diagram showing the frequency of different intestinal
deep endometriosis lesions according to Piketty’s et al.study [2].
sites represent up to 55% of the cases in the recent study by
Piketty et al.[2,3].The histologically proven rate of associa-
tion, in this study, with proximal ‘right’ intestinal (caecal or
ileal) lesions was 28% in patients with rectal and/or sigmoid
locations.
The usefulness of pre-treatment imaging
Different therapeutic strategies
Hormonal treatment of endometriosis can be effective on
the painful symptoms but does not improve fertility.It is the
first-line drug to patients with pain and no desire to become
pregnant. Surgical treatment is offered to patients whose
pain is not sufficiently improved by medical treatment and
also to patients who wish to become pregnant after two
IVF failures.Surgical treatment does indeed improve pain,
quality of life and fertility, provided that the lesions are
completely removed [4].
Diagnostic and Interventional Imaging (2013) 94, 281—291
282 A.Massein et al.
clinicopathological types of endometriosis can be distin-
guished, generally intricately linked: superficial periton-
eal endometriosis, ovarian endometriosis (cystic lesions
known as endometriomas) and deep pelvic endometriosis.
Extrapelvic locations (abdominal, pleural) are rare.
Some essential points concerning deep
pelvic endometriosis
There are two non-consensual definitions of deep pelvic
endometriosis: sub-peritoneal implants penetrating for
more than 5 mm under the peritoneum or infiltration of the
uterosacral ligaments and/or the muscles of adjacent pelvic
organs.
Deep pelvic endometriosis is responsible for chronic
pelvic pain (dysmenorrhea, deep dyspareunia), with or with-
out menstrual aggravation, but also dyschezia, rectorrhagia
and dysuria.It is also responsible for infertility with a fre-
quency estimated at 20 to 40% in cases of infertility.Rarely,
endometriosis is asymptomatic.
Its pathogenesis has not been elucidated. Several
hypotheses have been discussed including reflux of
endometrial fragments through the fallopian tubes dur-
ing menstruation, metaplasia of tissues derived from
the coelomic epithelium into endometrial tissue, vascular
and/or lymphatic emboli, and involvement of epigenetic and
environmental factors.
The diagnosis is often made too late, after an average
7 years of pain developing or after several years of medi-
cally assisted procreation treatment.Endometriosis can be
suspected clinically but imaging is the procedure that con-
firms it and maps the lesions. Laparoscopy should not be
performed for diagnostic purposes because it is not without
risk and can ignore some deep lesions or those masked by
adhesions.
Sites of deep endometriosis
Lesions of the uterosacral ligaments are the most common
(Fig.1).They are associated with involvement of the torus
uterinus, which is a small transverse thickening on the pos-
terior surface of the cervix, between the insertions of the
two uterosacral ligaments.
Intestinal and urinary tract locations are the most severe
forms of deep endometriosis (Fig.2).The multiple intestinal
Figure 1. Table showing the frequency of deep pelvic endometriosis according to Chapron’s surgical serie [3].
Figure 2. Diagram showing the frequency of different intestinal
deep endometriosis lesions according to Piketty’s et al.study [2].
sites represent up to 55% of the cases in the recent study by
Piketty et al.[2,3].The histologically proven rate of associa-
tion, in this study, with proximal ‘right’ intestinal (caecal or
ileal) lesions was 28% in patients with rectal and/or sigmoid
locations.
The usefulness of pre-treatment imaging
Different therapeutic strategies
Hormonal treatment of endometriosis can be effective on
the painful symptoms but does not improve fertility.It is the
first-line drug to patients with pain and no desire to become
pregnant. Surgical treatment is offered to patients whose
pain is not sufficiently improved by medical treatment and
also to patients who wish to become pregnant after two
IVF failures.Surgical treatment does indeed improve pain,
quality of life and fertility, provided that the lesions are
completely removed [4].
the ‘sliding sign’ in an anteverted uterus
• Gentle pressure is placed against the cervix using the
transvaginal probe, to establish whether the anterior rectum
glides freely across the posterior aspect of the cervix
(retrocervical region) and posterior vaginal wall.
• If the anterior rectal wall does so, the ‘sliding sign’ is
considered positive for this location
the ‘sliding sign’ in an anteverted uterus
The examiner then places one hand over the
woman’s lower anterior abdominal wall in order
to ballot the uterus between the palpating hand
and the transvaginal probe (which is held in the
other hand), to assess whether the anterior bowel
glides freely over the posterior aspect of the
upper uterus/fundus.
If it does so, the sliding sign is also considered
positive in this region
IDEA consensus opinion 321
(a) (b)
Figure 2 Schematic drawings demonstrating how to elicit the ‘sliding sign’ in an anteverted uterus (a) and a retroverted uterus (b).
If on TVS it is demonstrated that either the anterior
rectal wall or the anterior sigmoid wall does not glide
smoothly over the retrocervix or the posterior uterine
fundus, respectively, i.e. at least one of the locations
has a negative sliding sign, then the POD is recorded as
obliterated31,32
.
Demonstrating and describing the real-time
ultrasound-based sliding sign in a retroverted uterus is
different (Figure 2b). Gentle pressure is placed against
the posterior upper uterine fundus with the transvaginal
probe, to establish whether the anterior rectum glides
freely across the posterior upper uterine fundus. If the
anterior rectum does so, the sliding sign is considered
to be positive for this location (Videoclip S2a). The
between the palpating hand and transvaginal probe
anterior sigmoid glides freely over the anterior lower
uterine segment. If it does so, the sliding sign is also
considered to be positive in this region (Videoclip S2b).
As long as the sliding sign is found to be positive in both
of these anatomical regions (i.e. the posterior uterine
fundus and the anterior lower uterine segment), the POD
is recorded as non-obliterated33.
The fourth step is to search for DIE nodules in
the anterior and posterior compartments. To assess the
anterior compartment, the transducer is positioned in the
anterior fornix of the vagina. If bladder endometriosis
is suspected on the basis of symptoms, patients should
be asked not to empty their bladder completely before
the ultrasound examination. A slightly filled bladder
facilitates evaluation of the walls of the bladder and
detection and description of endometriotic nodules.
Finally, the transducer is positioned in the posterior fornix
of the vagina and slowly withdrawn through the vagina
to allow visualization of the posterior compartment.
Some authors advocate the use of bowel preparation
on the evening before the pelvic scan and the use of
a rectal enema within an hour before the ultrasound
examination to eliminate fecal residue and gas in the
rectosigmoid34–37. However, this is not mandatory, and
there are no published prospective studies comparing TVS
with and without bowel preparation for the diagnosis of
bowel DIE. In a recent meta-analysis, TVS, either with or
without bowel preparation, was found to be an accurate
predictor of rectosigmoid DIE8
.
COMPARTMENTAL EVALUATION
Anterior compartment
The anterior compartment includes the following anatom-
ical locations: urinary bladder, uterovesical region and
ureters.
Urinary bladder
Bladder DIE occurs more frequently in the bladder
base and bladder dome than in the extra-abdominal
bladder (Videoclip S3)38
. The bladder is best scanned
if it contains a small amount of urine because this
reduces false-negative findings. Although Savelli et al.38
described two zones (bladder base and dome), we propose
dividing the bladder ultrasound assessment into four
zones (Figure 3): (i) the trigonal zone, which lies within
3 cm of the urethral opening, is a smooth triangular
region delimited by the two ureteral orifices and the
internal urethral orifice (Figure S4a); (ii) the bladder base,
which faces backward and downward and lies adjacent
to both the vagina and the supravaginal cervix (Figure
S4b); (iii) the bladder dome, which lies superior to the
base and is intra-abdominal (Figure S4c); and (iv) the
extra-abdominal bladder (Figure S4d). Figure S5 and
Videoclip S3 demonstrate the most frequent location of
endometriotic bladder nodules, i.e. the bladder base.
On two-dimensional (2D) ultrasound the appearance
of DIE in the anterior compartment can be varied,
including hypoechoic linear or spherical lesions, with
or without regular contours involving the muscularis
the ‘sliding sign’ in a retroverted uterus
• Gentle pressure is placed against the posterior upper uterine
fundus with the transvaginal probe, to establish whether the
anterior rectum glides freely across the posterior upper
uterine fundus.
• If the anterior rectum does so, the sliding sign is considered
to be positive for this location
the ‘sliding sign’ in a retroverted uterus (b).
(a) (b)
Figure 2 Schematic drawings demonstrating how to elicit the ‘sliding sign’ in an anteverted uterus (a) and a retroverted uterus (b).
If on TVS it is demonstrated that either the anterior
rectal wall or the anterior sigmoid wall does not glide
smoothly over the retrocervix or the posterior uterine
fundus, respectively, i.e. at least one of the locations
has a negative sliding sign, then the POD is recorded as
obliterated31,32
.
Demonstrating and describing the real-time
ultrasound-based sliding sign in a retroverted uterus is
different (Figure 2b). Gentle pressure is placed against
the posterior upper uterine fundus with the transvaginal
probe, to establish whether the anterior rectum glides
freely across the posterior upper uterine fundus. If the
anterior rectum does so, the sliding sign is considered
to be positive for this location (Videoclip S2a). The
between the palpating hand and transvaginal probe
anterior sigmoid glides freely over the anterior lower
uterine segment. If it does so, the sliding sign is also
considered to be positive in this region (Videoclip S2b).
As long as the sliding sign is found to be positive in both
of these anatomical regions (i.e. the posterior uterine
fundus and the anterior lower uterine segment), the POD
is recorded as non-obliterated33.
The fourth step is to search for DIE nodules in
the anterior and posterior compartments. To assess the
anterior compartment, the transducer is positioned in the
anterior fornix of the vagina. If bladder endometriosis
is suspected on the basis of symptoms, patients should
examination to eliminate fecal residue and gas in the
rectosigmoid34–37. However, this is not mandatory, and
there are no published prospective studies comparing TVS
with and without bowel preparation for the diagnosis of
bowel DIE. In a recent meta-analysis, TVS, either with or
without bowel preparation, was found to be an accurate
predictor of rectosigmoid DIE8
.
COMPARTMENTAL EVALUATION
Anterior compartment
The anterior compartment includes the following anatom-
ical locations: urinary bladder, uterovesical region and
ureters.
Urinary bladder
Bladder DIE occurs more frequently in the bladder
base and bladder dome than in the extra-abdominal
bladder (Videoclip S3)38
. The bladder is best scanned
if it contains a small amount of urine because this
reduces false-negative findings. Although Savelli et al.38
described two zones (bladder base and dome), we propose
dividing the bladder ultrasound assessment into four
zones (Figure 3): (i) the trigonal zone, which lies within
3 cm of the urethral opening, is a smooth triangular
region delimited by the two ureteral orifices and the
internal urethral orifice (Figure S4a); (ii) the bladder base,
which faces backward and downward and lies adjacent
The examiner then places one hand over the
woman’s lower anterior abdominal wall in
order to ballot the uterus between the
palpating hand and transvaginal probe (which is
held in the other hand), to assess whether the
anterior sigmoid glides freely over the anterior
lower uterine segment.
If it does so, the sliding sign is also considered
to be positive in this region

8.02.2019
3
The fourth step is to search for DIE nodules in the
anterior and posterior compartments
• To assess the anterior compartment, the transducer is positioned
in the anterior fornix of the vagina.
• If bladder endometriosis is suspected on the basis of symptoms,
patients should be asked not to empty their bladder completely
before the ultrasound examination.
• A slightly filled bladder facilitates evaluation of the walls of the
bladder and detection and description of endometriotic nodules.
• Finally, the transducer is positioned in the posterior fornix of the
vagina and slowly withdrawn through the vagina to allow
visualization of the posterior compartment
proposed ultrasound definition of the
rectovaginal septum (RVS)
Septum
Rectum
Vagina
Cervix
(a)
Free fluid
Figure 4 Schematic drawing (a) and ultrasound image (b) demonstrating our proposed ultrasound definition of the rectovaginal septum
(RVS). (a) The RVS is denoted by the double-headed green arrow, below (anatomically) the blue line passing along the lower border of the
posterior lip of the cervix. The posterior vaginal fornix lies between the blue line and the red line (the latter passing along the caudal end of
the peritoneum of the lower margin of the rectouterine peritoneal pouch (cul-de-sac of Douglas)). (b) The upper delimitation of the RVS is
where the blue line passes along the lower border of the posterior lip of the cervix.
Free
fluid
Figure 5 Schematic drawings and ultrasound images demonstrating isolated deep infiltrating endometriosis in the rectovaginal septum
(RVS, ). The green ellipses encircle the endometriotic nodules in the RVS. , bowel wall; , vaginal wall.
Isolated deep infiltrating endometriosis in the
rectovaginal
septum
Septum
Rectum
Vagina
Cervix
(a)
Free fluid
Figure 4 Schematic drawing (a) and ultrasound image (b) demonstrating our proposed ultrasound definition of the rectovaginal septum
(RVS). (a) The RVS is denoted by the double-headed green arrow, below (anatomically) the blue line passing along the lower border of the
posterior lip of the cervix. The posterior vaginal fornix lies between the blue line and the red line (the latter passing along the caudal end of
the peritoneum of the lower margin of the rectouterine peritoneal pouch (cul-de-sac of Douglas)). (b) The upper delimitation of the RVS is
where the blue line passes along the lower border of the posterior lip of the cervix.
Free
fluid
Figure 5 Schematic drawings and ultrasound images demonstrating isolated deep infiltrating endometriosis in the rectovaginal septum
(RVS, ). The green ellipses encircle the endometriotic nodules in the RVS. , bowel wall; , vaginal wall.
RSV;
bowel wall;
vaginal wall.
deep infiltrating endometriosis in the posterior vaginal
wall with extension into the rectovaginal septum
Figure 6 Schematic drawings and ultrasound image demonstrating deep infiltrating endometriosis in the posterior vaginal wall with
extension into the rectovaginal septum (RVS, ). The green ellipses encircle the endometriotic nodules in the RVS. , bowel wall;
, vaginal wall.
Figure 7 Schematic drawings and ultrasound image demonstrating deep infiltrating endometriosis in the anterior rectal wall with extension
into the rectovaginal septum ( ). , bowel wall; , vaginal wall.
Figure 8 Schematic drawings and ultrasound image demonstrating rectovaginal septal ( ) deep infiltrating endometriosis with extension
into both anterior rectal wall and posterior vaginal wall. , bowel wall; , vaginal wall.
DIE54,58. The dimensions of the RVS DIE nodule should
be recorded in three orthogonal planes and the distance
between the lower margin of the lesion and the anal
verge should be measured. This should be done whether
the DIE is only in the vagina or only in the rectum,
or involves the vagina, RVS and rectum. Low RVS
lesions, when managed surgically, are associated with
severe complications, including fistulae56,59–61
.
Vaginal wall
We propose that involvement of the posterior vaginal
fornix and/or lateral vaginal fornix should be suspected
when a DIE nodule is seen on TVS in the rectovaginal
space below the line passing along the caudal end of
the peritoneum of the lower margin of the rectouterine
peritoneal pouch (cul-de-sac of Douglas) and above the
line passing along the lower border of the posterior lip of
the cervix (under the peritoneum) (seen in Figure 4).
Posterior vaginal fornix or forniceal endometriosis is
suspected if the posterior vaginal fornix is thickened or if
a discrete nodule is found in the hypoechoic layer of the
vaginal wall (Figure S8a). The hypoechoic nodule may be
homogeneous or inhomogeneous with or without large
cystic areas (Figure S8a) and there may or may not be
cystic areas surrounding the nodule6,39,41,42
. Figure S8b
is an ultrasound image demonstrating posterior vaginal
fornix DIE. The dimensions of the vaginal wall DIE nodule
should be measured in three orthogonal planes.
Rectovaginal nodules (‘diabolo’-like nodules)
Hourglass-shaped or ‘diabolo’-like nodules occur when
DIE lesions in the posterior vaginal fornix extend into
RSV; bowel wall; vaginal wall.
deep infiltrating endometriosis in the anterior rectal
wall with extension into the rectovaginal septum
, vaginal wall.
.
Vaginal wall
. Figure S8b
rectovaginal septal deep infiltratingendometriosis extending
intoboth anteriorrectal wall and posteriorvaginal wall.
, vaginal wall.
.
Vaginal wall
. Figure S8b

8.02.2019
4
ultrasound images of bowel deep infiltrating
endometriosis (DIE)IDEA consensus opinion 325
(a)
Transverse
section
Bowel
Uterus
Bowel
Bowel
Bowel
Bowel
(b)
(c)
(d)
(f)
(e)
(f)
Figure 9 Schematic drawings and corresponding ultrasound images of bowel deep infiltrating endometriosis (DIE). (a) DIE nodule with a
regular outline (absence of ‘spikes’). (b) DIE nodule with progressive narrowing, like a ‘tail’, also known as ‘comet’ sign. (c) DIE nodule with
prominent spikes towards the bowel lumen, also known as ‘Indian headdress’ or ‘moose antler’ sign. (d) DIE nodule with both prominent
spikes towards the bowel lumen (Indian headdress/moose antler sign) and progressive narrowing like a tail (comet sign). (e) DIE nodule with
both prominent spikes towards the bowel lumen (Indian headdress/moose antler sign) and extrinsic retraction (and visible mucosal folds)
(known as ‘pulling sleeve’ sign). The sliding sign is expected to be negative. (f) DIE nodule and extrinsic retraction (pulling sleeve sign). The
sliding sign is expected to be negative. Ultrasound image shows bowel adherent to the ovary; the hypoechogenic area between the bowel
loops and the ovary to which the bowel is adherent correspond to linear endometriotic changes (containing endometriotic glands and
stroma) not involving the muscularis of the bowel wall but located between the bowel and the ovary.
the anterior rectal wall62
(Figure S9a). On ultrasound,
the part of the DIE lesion situated in the anterior
rectal wall is the same size as the part situated in
the posterior vaginal fornix (Figure S9b). There is a
small but easily visualized continuum between these two
parts of the lesion. These lesions are located below the
peritoneum of the POD and are usually large (3 cm on
average)63
.
Uterosacral ligaments
Normal USLs are usually not visible on ultrasound (Figure
S10a). USL DIE lesions can be seen in the mid-sagittal view
of the uterus (Figure S10b). However, these are best seen
by placing the transvaginal probe in the posterior vagi-
nal fornix in the midline in the sagittal plane and then
sweeping the probe inferolaterally to the cervix. USLs
(a) DIE nodule with a
regular outline (absence
of ‘spikes’).
(b) DIE nodule with
progressive narrowing,
like a ‘tail’, also known as
‘comet’ sign.
endometriosis (DIE)
(c) DIE nodule with prominent spikes towards the bowel lumen, also known as
‘Indian headdress’ or ‘moose antler’ sign.
(a)
Transverse
section
Bowel
Uterus
Bowel
Bowel
Bowel
Bowel
(b)
(c)
(d)
(f)
(e)
(f)
average)63
.
endometriosis (DIE)
(a)
Transverse
section
Bowel
Uterus
Bowel
Bowel
Bowel
Bowel
(b)
(c)
(d)
(f)
(e)
(f)
average)63
.
(d) DIE nodule with both prominent spikes towards the bowel lumen (Indian headdress/moose antler
sign) and progressive narrowing like a tail (comet sign)
(e) DIE nodule with both prominent spikes towards the bowel lumen (Indian headdress/moose antler
sign) and extrinsic retraction (and visible mucosal folds) (known as ‘pulling sleeve’ sign)
endometriosis (DIE)
(a)
Transverse
section
Bowel
Uterus
Bowel
Bowel
Bowel
Bowel
(b)
(c)
(d)
(f)
(e)
(f)
average)63
.
(f) DIE nodule and extrinsic retraction
(pulling sleeve sign).
Rectum, rectosigmoid junction and sigmoidGuerriero and Condous et al.
red to be affected by DIE when a hypoechoic
with regular or irregular margins is seen within
eal fat surrounding the USLs. The lesion may be
may be part of a larger nodule extending into
or into other surrounding structures. The thick-
ickened’ USL can be measured in the transverse
e insertion of the ligament on the cervix pro-
he ligament can be distinguished clearly from
ructures (Figure S10c). In some cases the DIE
lving the USL is located at the torus uterinus
0d). If so, it is seen as a central thickening of
rvical area64
. The dimensions of the USL DIE
uld be recorded in three orthogonal planes.
ctosigmoid junction and sigmoid
E classically involves the anterior rectum,
d junction and/or sigmoid colon, all of which
ualized using TVS. Figure S11a demonstrates
c drawing of a DIE lesion within the upper
ctum. Bowel DIE can take the form of an
sion or can be multifocal (multiple lesions
e same segment) and/or multicentric (multiple
cting several bowel segments, i.e. small bowel,
, cecum, ileocecal junction and/or appendix)65
.
TVS can be used to visualize multifocal rectal
e S11b), there are no published data assessing
ance. Computed tomographic colonography
etic resonance imaging (MRI) can be used
e both multifocal and multicentric bowel
sis65
.
cally, bowel endometriosis is defined as the
endometrial glands and stroma in the bowel
ing at least the muscularis propria66, where
ably induces smooth-muscle hyperplasia and
his results in thickening of the bowel wall
narrowing of the bowel lumen. Normal rectal
s can be visualized on TVS: the anterior
sa is seen as a thin hyperechoic line; the
propria is hypoechoic, with the longitudinal
scle (outer) and circular smooth muscle (inner)
y a faint thin hyperechoic line; the submucosa
hogenic; and the mucosa is hypoechoic37,67
2a). Bowel DIE usually appears on TVS as a
of the hypoechoic muscularis propria or as
nodules, with or without hyperechoic foci
2b) with blurred margins. The morphological
wel lesion should be described according to
onographically, bowel lesions are hypoechoic
e cases a thinner section or a ‘tail’ is noted at
sembling a ‘comet’68
(Figure 9b). The normal
of the muscularis propria of the rectum
moid is replaced by a nodule of abnormal
possible retraction and adhesions, resulting
alled ‘Indian headdress’ or ‘moose antler’ sign
e,f)42
; the size of these lesions can vary.
1
2
3
4
Figure 10 Schematic drawing demonstrating distinction at
ultrasound between segments of the rectum and sigmoid colon for
specifying location of deep infiltrating endometriotic lesions: lower
(or retroperitoneal) anterior rectum (1); upper (visible at
laparoscopy) anterior rectum (2); rectosigmoid junction (3); and
anterior sigmoid (4).
located below the level of the insertion of the USLs
on the cervix being denoted as lower (retroperitoneal)
anterior rectal DIE lesions, those above this level being
denoted as upper (visible at laparoscopy) anterior rectal
DIE lesions, those at the level of the uterine fundus being
denoted as rectosigmoid junction DIE lesions and those
above the level of the uterine fundus being denoted as
anterior sigmoid DIE lesions (Figure 10). The dimensions
of the rectal and/or rectosigmoid DIE nodules should
between the lower margin of the most caudal lesion and
the anal verge should be measured using TVS.
Because bowel DIE may affect the bowel simultaneously
at different sites, other bowel lesions should be looked for
carefully when there is a DIE lesion affecting the rectum
(Figure S12b) or rectosigmoid. Preliminary data showed
that rectal DIE lesions may be associated with a second
intestinal lesion in 54.6% of cases34
.
Ultrasound diagnosis of POD obliteration31,32
has been
explained extensively earlier in this article. The oblitera-
tion can be graded as partial or complete depending on
whether one side (left or right) or both sides, respectively,
demonstrate a negative sliding sign. Furthermore, an
experienced operator can identify the level of POD oblit-
eration, i.e. specifying, in an anteverted uterus, whether
it is at the retrocervical level (lower third of the uterus),
mid-posterior uterus (middle third) and/or posterior uter-
ine fundus (upper third)69
and, in a retroverted uterus,
whether it is at the posterior uterine fundus, mid-anterior
uterus and/or lower anterior uterine wall33
(Figure S13).
MEASUREMENT OF LESIONS
lower (or retroperitoneal) anterior rectum (1);
upper (visible at laparoscopy) anterior rectum (2);
rectosigmoid junction (3);
and anterior sigmoid (4).
measurement of a nodule of deep infiltrating
endometriosis in the bowel wall.IDEA consensus opinion 327
Transverse
Sagittal
Figure 11 Schematic drawing and ultrasound images demonstrating measurement of a nodule of deep infiltrating endometriosis in the bowel
wall. Three orthogonal measurements should be taken, i.e. mid-sagittal, anteroposterior and transverse.
Figure 12 In cases of multifocal lesions of deep infiltrating endometriosis in the bowel, the total length of the bowel segment involved (from
caudal to cephalic aspect) should be measured, as shown in this schematic drawing and ultrasound image.
planes, to obtain the length (mid-sagittal measurement),
thickness (anteroposterior measurement) and transverse
diameter (Figure 11). This approach of measuring
in three planes applies to DIE lesions located in
the bladder, RVS, vagina, USLs, anterior rectum and
rectosigmoid.
Additionally, in cases of endometriosis in the ureters,
it is important to measure the distance between the
distal ureteric orifice and a DIE lesion which causes a
ureteric stricture; the stricture can be caused by either
extrinsic compression or intrinsic infiltration. Once the
and the lesion (Figure S15). It is possible to measure
the distance from the anus to the bowel lesion using
transrectal sonography. By inserting the probe into the
anus and positioning the tip of the probe up against
the endometriotic lesion71
, one’s finger can be kept on the
probe at the level of the anus and a ruler used to measure
the distance from the finger on the probe to the tip of
the probe when the probe has been withdrawn. TVS can
also be used to approximate the distance from the anal
verge to the lower margin of the bowel lesion. If there
are multifocal bowel lesions, then the distance between

8.02.2019
5
multifocal lesions of deep infiltratingendometriosis
total length of thebowel segment involved (caudal tocephalic)
Transverse
Sagittal
Figure 11 Schematic drawing and ultrasound images demonstrating measurement of a nodule of deep infiltrating endometriosis in the bowel
wall. Three orthogonal measurements should be taken, i.e. mid-sagittal, anteroposterior and transverse.
Figure 12 In cases of multifocal lesions of deep infiltrating endometriosis in the bowel, the total length of the bowel segment involved (from
caudal to cephalic aspect) should be measured, as shown in this schematic drawing and ultrasound image.
planes, to obtain the length (mid-sagittal measurement),
thickness (anteroposterior measurement) and transverse
diameter (Figure 11). This approach of measuring
in three planes applies to DIE lesions located in
the bladder, RVS, vagina, USLs, anterior rectum and
rectosigmoid.
Additionally, in cases of endometriosis in the ureters,
it is important to measure the distance between the
distal ureteric orifice and a DIE lesion which causes a
ureteric stricture; the stricture can be caused by either
extrinsic compression or intrinsic infiltration. Once the
stricture is identified along the longitudinal course of the
ureter, one caliper should be placed at this level and
the other at the distal ureteric orifice for measurement
(Figure S7).
In cases of multifocal bowel DIE lesions the total
mid-sagittal length of the bowel segment involved, from
caudal to cephalic aspect, should be measured (Figure 12).
It is important to be aware that the retraction within
rectosigmoid DIE lesions can result in an overestimation
of the true thickness of the lesion and an underestimation
of the true length of the lesion (Figure S14). This has
been described as the ‘mushroom cap’ sign on MRI and
can also be noted on TVS70
.
In cases of DIE lesions in the bowel or RVS, it is
important to measure the distance between the anal verge
and the lesion (Figure S15). It is possible to measure
the distance from the anus to the bowel lesion using
transrectal sonography. By inserting the probe into the
anus and positioning the tip of the probe up against
the endometriotic lesion71
, one’s finger can be kept on the
probe at the level of the anus and a ruler used to measure
the distance from the finger on the probe to the tip of
the probe when the probe has been withdrawn. TVS can
also be used to approximate the distance from the anal
verge to the lower margin of the bowel lesion. If there
are multifocal bowel lesions, then the distance between
the anal verge and the most caudal bowel lesion is
measured.
Figure 13 gives an overview of anterior and pos-
terior compartmental locations for deep infiltrating
endometriosis.
OTHER ULTRASOUND TECHNIQUES
Color Doppler
Although well established in the evaluation of
endometrioma20
, no prospective data have been reported
for the role of color Doppler in the evaluation of DIE.
Usually, endometriotic lesions in the rectosigmoid are
poorly vascularized. Color Doppler is useful in the dif-
ferential diagnosis between DIE in the bowel and rectal
Decision making for surgical approach
Roman H, Endometriosis Summit, Iceland 2017
Sonovaginography
been shown recently that medical treatment performed significantly better in relieving pain in women
without DIE compared with women with recto-vaginal lesions [46].
It seems, therefore, that as much detail as possible about the spread and localisation of the disease
are needed by the surgeon and clinician responsible for medical treatment. The careful evaluation of
clinical and diagnostic imaging findings gives clinicians the opportunity to decide the best surgical
approach, the possible need to involve surgical specialists other than a gynaecologic surgeon (e.g.
colorectal surgeon or urologist), so that management of the disease can be tailored correctly, and
patients can be informed of the extent of their disease and the therapeutic options available.
Ultrasound evaluation of deep infiltrating endometriosis
Transvaginal sonography can evaluate all potential locations of DIE in the anterior (bladder) or
posterior–lateral compartment. These include the rectovaginal septum, uterosacral ligaments, torus
uterinum (i.e. tissue behind the cervix in the mid-sagittal plane between the uterosacral ligaments),
posterior vaginal fornix, rectum and rectosigmoid junction, and parametria and ureteral involvement.
Endometriotic nodules of the bladder and the rectum can be evaluated with a transvaginal probe
and, if necessary, a transrectal examination with the same convex probe can be carried out. During the
transrectal examination, a fluid contrast medium can be inserted into the vagina (Fig. 6) to better
visualise the recto-vaginal septum.
Transabdominal ultrasound does not accurately detect DIE, mainly because bowel gas reduces the
ability to evaluate abdominal retroperitoneal or small bowel lesions, which are difficult to detect with
transabdominal ultrasound probes. Only endometriotic nodules of the abdominal wall can be easily
evaluated by a high-frequency transabdominal probe.
Deep nodes appear as hypoechoic lesions, linear or nodular retroperitoneal thickening with
irregular borders, and few vessels at power Doppler evaluation [47–50].
Women with suspected endometriomas associated with deep endometriosis, in particular those
with a frozen pelvis or recto-vaginal or bladder nodules, should fist undergo a detailed examination of
the pelvis to evaluate the anatomy of the uterus and the adnexa, both in the sagittal and horizontal
plane, with gentle probe movements to assess the presence of adhesion between them. Transvaginal
sonographic examination is based on a detailed evaluation of organ and tissues dividing the pelvis in
the anterior and posterior compartment according to the DIE classification by Chapron et al. [51].
Fig. 6. Sonovaginography: visualisation of the vagina with transvaginal probe positioned in the rectum. The vagina is filled with
saline solution through a Foley catheter with its balloon placed in the lower part of the vagina; note the wall of the vaginal posterior
fornix, the cervix in the vagina, the recto-vaginal septum, and the retrocervical nodule of deep infiltrating endometriosis not
invading the vagina but infiltrating the lower part of the rectal wall and the left uterosacral sacral ligament.
Visualisation of the vagina with
transvaginal probe positioned in
the rectum.
The vagina is filled with saline
solution through a Foley catheter
with its balloon placed in the
lower part of the vagina
a dilated ureter seen by transvaginal ultrasound
Vagina. In the vagina, nodular thickening of the posterior vaginal fornix takes place, which does not
become thinner with probe compression. The insertion of saline solution into the vagina (sonovagi-
nography) can improve visualisation of these lesions [53]. (Fig. 6) An increase in the amount of ul-
trasonographic gel inside the probe’s cover can also improve visualisation of the vaginal walls and
posterior and anterior fornix [52]. The low reported accuracy of transvaginal sonography in detecting
vaginal endometriosis [36,47] confirms that digital gynaecological examination could be a better
alternative to transvaginal sonography. In the recto-vaginal septum, nodules replacing the normal
hyperechoic aspect of the tissue between the vagina and the rectum are present below the horizontal
plane passing through the lower border of the posterior lip of the cervix [47].
In the torus, a nodular image with irregular margins behind the cervix is found in the mid-sagittal
plane. The endometriotic lesion in uterosacral ligaments is visible near the insertion on the posterior
lateral cervix wall, as a nodule with regular or stellate margins or as hypoechoic linear thickening.
In cases of endometriotic lesions involving the uterosacral ligament, special attention must be paid
to the parametrium and to the pelvic ureteral evaluation, particularly in the paracervical area. Para-
metria are examined lateral to the uterine cervix firstly on the sagittal planes moving the probe from
the lateral sites where the parametrium is attached to the cervix, to the uterine vessels bifurcation, to
the lateral pelvic wall, and then on the transverse planes moving the probe from the uterine isthmus to
the external cervical orifice. Parametrial involvement is seen as an infiltrating hypoechogenic irregular
tissue, and can be medially delimited from the cervical vascular plexuses using colour or power
Doppler.
Ureter. Pelvic ureteral dilatation can be easily seen by transvaginal sonography as a tubular anechoic
image (Fig. 8) with or without movements in the parametrial tissue, similar to a blood vessel but with
negative Colour or power Doppler signs. In the case of extrinsic compression without stenosis of the
ureter, the transvaginal sonographic diagnosis is more difficult. The distal part of the ureter can be
identified adjacent to the bladder trigon, and followed laterally to the cervix, to the pelvic brim and to
the level where it crosses the common iliac vessels [54]. An extrinsic compression also without ureteral
dilatation could be suspected in cases in which a DIE lesion is located close to the ureter. The hypothesis
of ureteral involvement suggests a specific and accurate evaluation at the time of surgery, and, in these
cases, transabdominal ultrasound to evaluate the renal pelvis should be added.
Rectum and recto-sigmoid junction. Rectal sigmoid nodules are visualised as an irregular hypoechoic
mass penetrating into the intestinal wall distorting its normal structure At transvaginal sonography,
Fig. 8. Ultrasound image of a dilated ureter seen by transvaginal ultrasound in a transverse section of the pelvis, as tubular anechoic
structure (arrows) in the parametrial tissue laterally to the uterine cervix.
In a transverse section of the pelvis,
seen as tubular anechoic structure
(arrows) in the parametrial tissue
laterally to the uterine cervix.
Water-contrast in the rectum during
transvaginal sonography
woman’s lower anterior abdominal wall to ballot the uterus between the palpating hand and trans-
vaginal probe to determine whether the recto sigmoid glides freely over the posterior aspect of the
upper uterus or fundus [38,39].
Recent studies have shown that transvaginal sonography, when carried out by experienced
sonographers, may indeed be a highly valuable test for the detection of DIE [47–50,56,57].
The reported accuracy of the ultrasound diagnosis of DIE varies between studies, which may reflect
the variations in the examination technique, quality of ultrasound equipment, and experience of the
operators. The prevalence of disease is also variable in different studies, which may bias the findings.
Although reported sensitivity and specificity of transvaginal sonography in the prediction of DIE is
high [47–58], evaluation of DIE by transvaginal sonography is difficult and requires expertise. There-
fore, some easily detectable utrasonographic signs have recently been proposed to predict the risk of
the presence of DIE. Real-time dynamic transvaginal sonography evaluation of the posterior
compartment using the ‘sliding sign’ seems to establish whether the pouch of Douglas is obliterated,
and may also be useful in the identification of women who may be at a higher risk for bowel endo-
metriosis [38,39].
Transvaginal sonography has low accuracy in diagnosing the infiltration of the mucosal layer [4].
Also, transrectal ultrasound, which is a valuable tool for detecting rectal endometriosis as endo-
metriotic infiltration of the muscularis layer, is less accurate in assessing submucosal, mucosal layer
involvement, or both [49,59]. Therefore, transvaginal and transrectal sonography does not help sur-
geons in deciding whether or not to perform segmental or discoid resection of the lesion. More likely,
this decision depends on the patient’s symptoms and is also related to the diameters of infiltrating
tissue and the presence of lumen stenosis. It has been reported that adding water-contrast in the
rectum during transvaginal ultrasonography improves the diagnosis of rectal infiltration in women
with rectovaginal endometriosis [60]. Saline solution is injected into the rectal lumen under ultraso-
nographic control through a catheter (Fig.11). Presence of rectovaginal nodules, presence and degree of
rectal infiltration, and the largest diameter of the bowel nodules can be evaluated. The procedure
determines the presence of rectovaginal nodules infiltrating the rectal muscularis propria more
accurately than transvaginal sonography [60]. It can be used when transvaginal sonography cannot
exclude the presence of rectal infiltration. In cases of suspected bowel stenosis based on symptoms and
on transvaginal sonography findings, a barium enema could help decide whether segmental resection
is necessary.
Fig. 11. Water-contrast in the rectum during transvaginal sonography is performed by injecting saline solution into the rectal lumen
during transvaginal ultrasound examination. Note the presence of the deep infiltrating endometriosis nodule bulging into the bowel
lumen. The lesion clearly reduces the rectal lumen. It infiltrates only the muscle layer of the bowel. The lesion is covered by the
hyperechogenic submucosa and hypoechogenic mucosa.
injecting saline
solution into the
rectal lumen
during
transvaginal
ultrasound
examination
Water enema CT288 A. Massein et al.
Figure 8. Water enema CT showing typical retrocervical involvement: nodule (arrow) of the anterior wall of the upper rectum adhering
to the torus uterinus. a: axial slice. b: sagittal slice.
Figure 9. Water enema CT showing parietal thickening of the sigmoid colon (arrow) contiguous with a left ovarian lesion (star). a: axial
slice. b: sagittal oblique reconstruction.
Typical retrocervical
involvement:
nodule (arrow) of the
anterior wall of the
upper rectum adhering
to the torus uterinus.
a: axial slice.
b: sagittal slice.

8.02.2019
6
Water enema CT
288 A. Massein et al.
Figure 8. Water enema CT showing typical retrocervical involvement: nodule (arrow) of the anterior wall of the upper rectum adhering
to the torus uterinus. a: axial slice. b: sagittal slice.
Figure 9. Water enema CT showing parietal thickening of the sigmoid colon (arrow) contiguous with a left ovarian lesion (star). a: axial
slice. b: sagittal oblique reconstruction.
parietal thickening of the
sigmoid colon (arrow)
contiguous with a left
ovarian lesion (star).
a: axial slice.
b: sagittal oblique
reconstruction.
Water enema CTImaging of intestinal involvement in endometriosis 289
Figure 10. Water enema CT showing four sites of intestinal endometriosis in the same patient. a:nodular thickening of the torus uterinus
adhering to the anterior surface of the upper rectum. b, c: Thickening of the wall of the sigmoid colon (arrow) in axial (b) and oblique
sagittal (c) slices adhering to the torus uterinus and the left ovary (site of several endometrioma detected by ultrasound). d, e: rounded
nodule in the wall of the final loop of the ileum (arrow) in axial (d) and coronal (e) slices. f: parietal thickening of an ileal loop (arrow) in
contact with the right ovary which is also involved.
a: nodular thickening of the torus uterinus adhering to the anterior surface of the
upper rectum.
b, c: Thickening of the wall of the sigmoid colon (arrow) in axial (b) and oblique
sagittal (c) slices adhering to the torus uterinus and the left ovary (site of several
endometrioma detected by ultrasound).
Imaging of intestinal involvement in endometriosis
Figure 10. Water enema CT showing four sites of intestinal endometriosis in the same patient. a:nodular thickening
adhering to the anterior surface of the upper rectum. b, c: Thickening of the wall of the sigmoid colon (arrow) in a
sagittal (c) slices adhering to the torus uterinus and the left ovary (site of several endometrioma detected by ultraso
nodule in the wall of the final loop of the ileum (arrow) in axial (d) and coronal (e) slices. f: parietal thickening of an
Water enema CT
estinal involvement in endometriosis 289
ater enema CT showing four sites of intestinal endometriosis in the same patient. a:nodular thickening of the torus uterinus
anterior surface of the upper rectum. b, c: Thickening of the wall of the sigmoid colon (arrow) in axial (b) and oblique
s adhering to the torus uterinus and the left ovary (site of several endometrioma detected by ultrasound). d, e: rounded
all of the final loop of the ileum (arrow) in axial (d) and coronal (e) slices. f: parietal thickening of an ileal loop (arrow) in
e right ovary which is also involved.
e: rounded nodule in the wall of the final loop of the ileum (arrow) in
axial (d) and coronal (e) slices. f: parietal thickening of an ileal loop
(arrow) in contact with the right ovary which is also involved.
Imaging of intestinal involvement in endometriosis 289
Figure 10. Water enema CT showing four sites of intestinal endometriosis in the same patient. a:nodular thickening of the torus uterinus
adhering to the anterior surface of the upper rectum. b, c: Thickening of the wall of the sigmoid colon (arrow) in axial (b) and oblique
sagittal (c) slices adhering to the torus uterinus and the left ovary (site of several endometrioma detected by ultrasound). d, e: rounded
nodule in the wall of the final loop of the ileum (arrow) in axial (d) and coronal (e) slices. f: parietal thickening of an ileal loop (arrow) in
Muscularis involvement at the rectosigmoid junction
290 A. Massein et al.
Figure 11. Rectal ES - axial section. Nodular hypoechoic thick-
ening of the muscularis (arrow). Submucosa (thin, hyperechoic
appearance [arrowhead]) spared. Normal appearance (hypoechoic
and fine) of the muscularis (star).
Figure 12. Muscularis involvement at the rectosigmoid junction, with correlation in transvaginal ultrasound (a), MRI (b), Enema CT (c),
rectal ES (d). a:sagittal slice showing hypoechoic nodular thickening of the muscularis (arrow). b:T2-weighted axial slice showing an arcuate
hypointense thickening of the uterosacral ligaments. On the right, this thickening is adhering to an intestinal parietal nodule. c: oblique
sagittal reconstruction showing a single intestinal location. d: nodule in the muscularis layer, sparing the submucosa, located 20 cm from
the anal margin.
transvaginal ultrasound (a),
MRI (b),
Enema CT (c),
rectal ES (d)
MR imaging
• The state-of-the-art MR imaging protocol for the diagnosis of
endometriosis includes T2- and fat suppressed T1-weighted
sequences.
• T2-weighted sequences without fat-suppression are the best
sequences for detecting pelvic endometriosis, in particular for
the evaluation of fibrotic lesions.
• Fat-suppressed T1-weighted 3D gradient-echo LAVA sequence.
This pulse sequence improves the sensitivity of MR imaging in
the detection of small lesions. It is the most sensitive for the
detection of bloody foci and peritoneal endometriosis
MR imaging
• Contrast-enhanced fat-suppressed T1-weighted 3D gradient-
echo LAVA sequence is useful in the following conditions:
• detection of enhancing mural nodules within adnexal masses, when
atypical features on US or T2-weighted MR sequences suggest potential
malignancy
• the major benefit of intravenous gadolinium is ureter visualization

Diagnosis of Endometriosis

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Diagnosis of Endometriosis

Similar to Diagnosis of Endometriosis (20)

More from Tevfik Yoldemir

More from Tevfik Yoldemir (20)

Recently uploaded

Recently uploaded (20)

Diagnosis of Endometriosis