Keywords lesion size - preoperative evaluation - #Enzian classification - transvaginal sonography
(TVS) - THEMES - surgery
Introduction
Endometriosis may manifest in different forms, with superficial peritoneal lesions
representing the mildest form of the disease. Endometriotic cysts of the ovaries are
frequent findings and varying degrees of associated adhesions may also be encountered.
Deep endometriosis (DE) may affect the urinary bladder, rectum, vagina, rectovaginal
septum, uterosacral ligaments (USLs), or parametria [1 ]
[2 ]
[3 ]
[4 ]. Various classification systems have been developed to describe the extent of the
disease. Many of them are based on an intraoperative evaluation of peritoneal disease
location and severity. In contrast, the Enzian classification, which was originally
created in 2003 in an attempt to better describe the location and severity of DE [5 ], until now focused on deep disease. However, efforts have been made to create an
updated, universally usable and applicable classification – the #Enzian classification,
which was published in January, 2021 and now also comprises the assessment of endometriotic
lesions of the ovaries and the peritoneum as well as adhesions of the ovaries and
tubes in addition to DE [6 ]. As a consequence, it may enable a more comprehensive representation of the extent
of the disease. The #Enzian classification was primarily developed as a surgical score,
but the focus was also put on the possibility to additionally apply it to the findings
of different imaging techniques such as transvaginal sonography (TVS) and magnetic
resonance imaging (MRI). This is particularly important in light of the fact that
surgical planning may be facilitated and conservative treatment monitoring may also
be made easier with a uniform language. The #Enzian classification aims to evaluate
all #Enzian compartments also by TVS or MRI except for compartment P, which describes
peritoneal lesions that can hardly be detected by imaging [6 ]. The aim of the present study was to exactly compare endometriotic lesion sizes
evaluated by preoperative TVS and surgery for all #Enzian compartments in which a
differentiation of specific severity grades is foreseen by this classification. Furthermore,
the severity grades of adhesions of the ovaries and tubes were compared between TVS
and surgery, and the sensitivities and specificities for the detection of endometriotic
lesions and adhesions by TVS in the different #Enzian compartments were calculated.
Methods
This study is a retrospective data analysis of all women aged 18 years or older who
underwent preoperative TVS evaluation followed by surgical treatment for DE at a tertiary
referral center for endometriosis in Italy between January 1, 2019 and December 31,
2019. The study was approved by the local IRB and the STARD guidelines were followed
for this study. Only patients with TVS performed within three months before the surgical
intervention were included. Patients with diagnosed or suspected malignancy, a previous
colorectal surgery, or a previous surgery for DE including vaginal resection, full
thickness bowel resection, or excision of a DE lesion involving the urinary bladder
were excluded. TVS findings regarding endometriotic lesions were extracted from medical
records. In this regard, the location and lesion size of every endometriotic lesion
were described following the criteria of the #Enzian classification [6 ]. Furthermore, the exact description of lesion size in mm was extracted. All preoperative
sonographic examinations had been carried out by a gynecologist with extensive ultrasound
experience especially in the field of endometriosis (A.D.G.) using a Voluson E8 ultrasound
device (GE Healthcare Austria GmbH, Vienna, Austria). Routinely, transabdominal sonography
had been performed in addition to TVS, for example to evaluate the kidneys with regard
to hydronephrosis in the presence of DE or to assess the presence of endometriotic
lesions in women with upper abdominal pain. As for the sonographic examinations, the
intraoperative findings regarding endometriotic lesions were extracted from the surgery
reports and the location and size of each lesion were described following the criteria
of the #Enzian classification [6 ]. Furthermore, the description of lesion size in mm was extracted. In this regard,
all lesions had been described in 5 mm increments at surgery (i.e., lesion size of
5 mm, 10 mm, 15 mm, etc.). All surgeries had been performed by a pelvic surgeon and
his team of gynecologic surgeons with extensive experience in the field of minimally
invasive surgery, especially for DE, including urologic and colorectal procedures.
The surgeons were not blinded regarding the preoperative TVS examination, as the TVS
examinations are routinely performed for the planning of the surgery.
For endometriotic lesions involving #Enzian compartment A (vagina, retrocervical area
and/or rectovaginal septum; [Fig. 1 ], panel a), B (uterosacral ligaments, parametrium; [Fig. 1 ], panel b), and C (rectum; [Fig. 1 ], panel c), the definition proposed by the International Deep Endometriosis Analysis
group (IDEA) [7 ] statement was used regarding the vagina and rectovaginal septum. According to the
#Enzian classification, the size of an endometriotic lesion in #Enzian compartments
A, B, and C was described as severity stage 1=lesion size < 1 cm; stage 2=lesion size
1–3 cm, and stage 3=lesion size > 3 cm [6 ].
Fig. 1 Representative ultrasound images of endometriotic lesions in #Enzian compartment A
(panel a ), #Enzian compartment B (panel b ), #Enzian compartment C (panel c ), and #Enzian compartment O (panel d ).
For #Enzian compartment O (ovary; [Fig. 1 ], panel d), each side (left/right) was considered separately and the different severity
stages of ovarian endometriotic cysts on the respective side were classified as stage
1 = maximal diameter of the lesion (or sum of the maximal diameters of all cysts in
the presence of more than one cyst in an ovary) < 3 cm; stage 2 = 3–7 cm and stage
3 = > 7 cm. Furthermore, missing ovaries on one or both sides were recorded.
#Enzian compartment T (tubo-ovarian condition) was again evaluated for each side separately
(left/right) and the severity stages were extracted from the TVS and surgery reports
using the following criteria: stage 1 = mild adhesions (tube, ovary, and pelvic side
wall appearing to adhere to each other); stage 2 = moderate adhesions (additional
adhesions to the uterus); stage 3 = severe (adhesions additionally affecting the bowel
and/or USL).
Furthermore, DE lesions affecting #Enzian compartment FB (urinary bladder; [Fig. 2 ], panel a), FI (other intestinal locations, i.e., sigmoid colon, small bowel, etc.),
FU (ureters; [Fig. 2 ], panel b) and FO (other extragenital lesions; [Fig. 2 ], panel c) were evaluated.
Fig. 2 Representative ultrasound images of endometriotic lesions in #Enzian compartment FB
(panel a ), #Enzian compartment FU (panel b ), and #Enzian compartment FO (in this case a lesion of the abdominal wall, panel
c ).
#Enzian compartment P (peritoneal lesions) was not evaluated in this study, as it
can only be assessed during surgery but not during TVS. Furthermore, due to the retrospective
approach of the present study, this information was not available from surgical reports
according to the #Enzian criteria. #Enzian compartment FA (adenomyosis) was also not
assessed because definitive histological diagnosis regarding adenomyosis was not available
for all cases.
Endometriosis was confirmed histologically in all included cases. Different patient
characteristics such as age, body mass index, gravidity, parity, as well as data on
preoperative pain symptoms such as dysmenorrhea, dyspareunia, dysuria, dyschezia,
infertility, constipation, diarrhea, and rectal bleeding were extracted from the patients'
charts.
Statistical Analysis
Data were represented by descriptive statistics. For #Enzian compartments A, B, C,
and O, the measurements in mm were recorded in addition to the severity grades of
the #Enzian classification. For TVS, they were given as the sizes measured in 1 mm
increments during the examination, whereas in the case of surgery, a description of
the lesion sizes in 5 mm increments was recorded. Therefore, if a lesion was recorded
on TVS, for example as 10 mm, the real size ranged from 9.5 to 10.5 mm. If a lesion
was recorded during surgery as 10 mm, the real size ranged between 7.5 and 12.5 mm.
Starting from these considerations, the percentages of concordance (with 95% confidence
intervals) between TVS and surgery were calculated by assessing the number of cases
in which the size interval according to TVS was entirely contained in the size interval
according to surgery. In this regard, for #Enzian compartments B and O, the left and
the right side were evaluated and described separately. For #Enzian compartment T,
the findings during TVS and surgery were compared using the #Enzian severity grades.
In addition to the exact concordances of the lesion sizes in mm between TVS and surgery,
the rates of concordance between the lesion sizes described according to the severity
grades of the #Enzian classification were also calculated for compartments A, B, C,
and O.
Furthermore, the overall sensitivities and specificities for the detection of endometriotic
lesions (or adhesions, respectively) in #Enzian compartments A, B, C, O, FB, FU, FI,
and FO by TVS were calculated. Again, for #Enzian compartments B and O, the left and
the right sides were considered separately, and the sensitivities and specificities
were calculated for each side.
Results
Between January 1, 2019 and December 31, 2019, 93 women who had undergone preoperative
TVS and surgical resection of DE were included. The patient characteristics of these
women are shown in [Table 1 ]. [Table 2 ] summarizes the number of women affected by endometriotic lesions/adhesions in each
#Enzian compartment including the number of cases for each severity grade in the respective
compartments.
Table 1 Patient characteristics (n = 93).
Patient characteristics
Value
a SD – standard deviation; b BMI – body mass index
Age [years], mean ± SDa
37.3 ± 6.6
BMI
b [kg/m²], mean ± SDa
23.7 ± 3.9
Gravidity
0, %
62
1, %
23
2, %
11
3, %
3
Parity
Nulliparity, %
69
Primiparity, %
18
Parity of 2, %
12
Parity of 3, %
1
Preoperative symptom
Dysmenorrhea, %
95
Dyspareunia, %
96
Dyschezia, %
88
Dysuria, %
32
Infertility, %
43
Constipation, %
75
Diarrhea, %
33
Rectal bleeding, %
5
Table 2 Affected #Enzian compartments as assessed by surgery.
Affected #Enzian compartment
Number of cases
a RVS – rectovaginal septum; b USLs – uterosacral ligaments
#Enzian A (vagina, retrocervical area, RVSa )
72
A1
3
A2
67
A3
2
#Enzian B left (left USLsb , left parametria)
70
B1
8
B2
58
B3
4
#Enzian B right (right USLsb , right parametria)
71
B1
12
B2
56
B3
3
#Enzian C (rectum)
44
C1
0
C2
17
C3
27
#Enzian O left (left ovary)
39
O1
14
O2
21
O3
4
Missing
0
#Enzian O right (right ovary)
31
O1
10
O2
15
O3
6
Missing
6
#Enzian T left (left tubo-ovarian condition)
88
T1
11
T2
25
T3
52
Missing
3
#Enzian T right (right tubo-ovarian condition)
82
T1
16
T2
23
T3
43
Missing
5
#Enzian FB (urinary bladder)
7
#Enzian FU (ureters)
13
#Enzian FI (other intestinal locations)
17
[Fig. 3 ] shows the percentages of cases (with 95% confidence intervals (CIs)) in which the
lesion size measurement by TVS matches exactly with the lesion size seen during surgery.
The highest rate of exact concordance was found for #Enzian compartment C, with exact
matches in 74% of cases ([Fig. 3 ], panel b). When considering a tolerance margin of a maximum of 3 mm for the TVS
measurements, 87% concordant cases were seen. Conversely, the lowest rate of exact
concordance was found for #Enzian compartment B left with 55% ([Fig. 3 ], panel c). However, when considering a maximum tolerance margin of 3 mm, a rate
of 88% concordant cases was seen also in this case. In #Enzian compartments A, C,
B left and B right, almost 100% concordance between TVS and surgery was reached when
a maximum tolerance margin of 10 mm (1 cm) was considered. For #Enzian compartment
O (ovaries), exact concordance rates of 65% on the left side and 72% on the right
were found ([Fig. 3 ], panels e and f). With a maximum tolerance margin of 10 mm, concordance rates of
90% on the left side and 92% on the right side were reached and with a maximum tolerance
margin of 20 mm, 97% on each side was achieved.
Fig. 3 Percentages with respective 95% confidence intervals of concordance between transvaginal
sonography (TVS) measurements and surgery regarding endometriotic lesions in different
#Enzian compartments. Exact concordance is given (i.e., the size interval according
to TVS being entirely contained in the size interval according to surgery, as outlined
in the Methods section), as well as concordance with a tolerance margin for TVS measurements
of a maximum of 1 mm, 3 mm, 5 mm and 10 mm (as well as 20 mm for the ovaries, panels
e and f ), respectively. Panel a : #Enzian compartment A (vagina, retrocervical area, rectovaginal septum). Panel b : #Enzian compartment C (rectum). Panel c : #Enzian compartment B left (left uterosacral ligament, left parametria). Panel d : #Enzian compartment B right (right uterosacral ligament, right parametria). Panel
e : #Enzian compartment O left (left ovary). Panel f : #Enzian compartment O right (right ovary).
For #Enzian compartment T (tubo-ovarian condition), reflecting adhesions of the ovary
and tube to the pelvic side wall and – depending on the severity grade – also to the
uterus or the bowel and/or USL, the comparisons between the findings during TVS and
surgery according to the #Enzian severity grades are shown in [Table 3 ]. The highest concordances in this compartment were seen for severe adhesions: severe
adhesions seen on TVS (i.e., TVS T3) were confirmed during surgery in 86% cases for
the left side and in 84% on the right side. In the remaining cases of TVS T3, the
adhesions were assessed as moderate (i.e., T2) during surgery. None of the TVS T3
assessments turned out to be only slight adhesions (T1) or even cases without any
adhesions (T0). Moderate adhesions on TVS (TVS T2) were classified as T2 in 68% (left
side) and 59% (right side) of cases during surgery, whereas the remaining cases were
classified as either slight or severe on both sides. Slight adhesions on TVS (T1)
were confirmed as such during surgery in 75% (left side) and 77% (right side) of cases,
whereas for the remaining TVS T1 cases, no adhesions were found during surgery (T0).
None of the TVS T1 cases turned out to be moderate (T2) or severe (T3) adhesions during
surgery.
Table 3 #Enzian severity grades of adhesions in #Enzian compartment T (tubo-ovarian condition)
according to preoperative transvaginal sonography and surgery.
Surgery
Transvaginal sonography (TVS)
Left side
T0
T1
T2
T3
T0 , % of TVS T0
0
0
0
0
T1 , % of TVS T1
25
75
0
0
T2 , % of TVS T2
0
20
68
12
T3, % of TVS T3
0
0
14
86
Right side
T0
T1
T2
T3
T0 , % of TVS T0
100
0
0
0
T1 , % of TVS T1
23
77
0
0
T2 , % of TVS T2
0
22
59
19
T3 , % of TVS T3
0
0
16
84
[Table 4 ] shows the concordance rates between the lesion sizes during TVS and surgery evaluated
according to the #Enzian severity grades for compartment A, B (left/right), C, and
O (left/right). The concordance rates for the assignment to the specific severity
grade ranged between 73% and 100%, except for A1 and B1 lesions, where only 29% to
46% of the A1 and B1 lesions on TVS were confirmed as such during surgery. The other
TVS A1 and B1 lesions were classified as lesions in adjacent severity grades ([Table 4 ]).
Table 4 #Enzian severity grades of lesions in #Enzian compartments A, B (left/right), C and
O (left/right) according to preoperative transvaginal sonography and surgery.
Surgery
Transvaginal sonography (TVS)
A0
A1
A2
A3
A0 , % of TVS A0
90
5
5
0
A1 , % of TVS A1
14
29
57
0
A2 , % of TVS A2
3
0
97
0
A3 , % of TVS A3
0
0
50
50
Left side
B0
B1
B2
B3
B0 , % of TVS B0
89
0
11
0
B1 , % of TVS B1
37
37
26
0
B2 , % of TVS B2
0
2
96
2
B3 , % of TVS B3
0
0
0
100
Right side
B0
B1
B2
B3
B0 , % of TVS B0
100
0
0
0
B1 , % of TVS B1
13
46
42
0
B2 , % of TVS B2
0
2
98
0
B3 , % of TVS B3
0
0
25
75
C0
C1
C2
C3
C0 , % of TVS C0
100
0
0
0
C1 , % of TVS C1
0
0
0
0
C2 , % of TVS C2
11
0
89
0
C3 , % of TVS C3
0
0
4
96
Left side
O0
O1
O2
O3
O0 , % of TVS O0
100
0
0
0
O1 , % of TVS O1
13
80
7
0
O2 , % of TVS O2
0
10
90
0
O3 , % of TVS O3
0
0
20
80
Right side
O0
O1
O2
O3
O0 , % of TVS O0
100
0
0
0
O1 , % of TVS O1
0
73
27
0
O2 , % of TVS O2
0
17
83
0
O3 , % of TVS O3
0
0
25
75
Confirmation rates by surgery (not sensitivities and specificities, but percentages
of positives on sonography of all true positives as well as percentages of negatives
on sonography of all true negatives) of the presence or absence of an endometriotic
lesion during sonographic examination in different #Enzian compartments were as follows:
FB (urinary bladder): 100% (95% CI 61% – 100%); 99% (95% CI 94% – 100%).
FI (other intestinal locations): 100% (95% CI 81% – 100%); 100% (95% CI 95% – 100%).
FU (ureters): 100% (95% CI 77% – 100%); 100% (95% CI 95% – 100%).
FO (other extragenital locations): 71.4% (95% CI 36% – 92%); 100% (95% CI 96% – 100%).
In summary, one DE lesion of the urinary bladder was not detected by TVS and the only
two discordant findings in #Enzian compartment FO were a lesion at the level of the
umbilicus and another one in the left broad ligament seen during the sonographic examination
which could not be confirmed during surgery. All other sonographic findings in #Enzian
compartments FB, FI, FU, and FO were correct.
The sensitivities and specificities for the detection of the presence of endometriotic
lesions by preoperative TVS in each #Enzian compartment with the respective 95% CIs,
considering only the presence or absence of a lesion irrespective of the severity
grade of the lesion, are shown in [Table 5 ]. For #Enzian compartment T, it was not possible to calculate sensitivities and specificities,
as there were too few cases without any lesion during TVS or surgery (there were only
2 women without any adhesions in compartment B left and 6 women without adhesions
in compartment B right; all other women presented with adhesions of varying severity
grades).
Table 5 Sensitivities and specificities for the detection of endometriotic lesions in different
#Enzian compartments by preoperative sonographic examination.
#Enzian compartment
Sensitivity (95% CIa )
Specificity (95% CIa )
a CI – confidence interval; b RVS – rectovaginal septum, c USLs – uterosacral ligaments
#Enzian FB (urinary bladder)
86% (42%-100%)
100% (96%-100%)
#Enzian FI (other intestinal locations)
100% (80%-100%)
100% (95%-100%)
#Enzian FU (ureters)
100% (75%-100%)
100% (95%-100%)
#Enzian FO (other location)
100% (48%-100%)
98% (92%-100%)
#Enzian A (vagina, retrocervical area, RVSb )
97% (90%-100%)
86% (64%-97%)
#Enzian B left (left USLsc , left parametria)
97% (90%-100%)
70% (47%-87%)
#Enzian B right (right USLsc , right parametria)
100% (95%-100%)
90% (70%-99%)
#Enzian C (rectum)
100% (92%-100%)
96% (86%-100%)
#Enzian O left (left ovary)
100% (91%-100%)
96% (87%-100%)
#Enzian O right (right ovary)
100% (87%-100%)
100% (94%-100%)
Discussion
This study shows that preoperative TVS can accurately predict the location and size
of DE lesions. In #Enzian compartments A, B, and C, more than 80% of all cases lay
within a maximum tolerance margin of 3 mm for the TVS measurements, reflecting a clinically
negligible difference in particular for large DE lesions. Confirming previous findings
[8 ], #Enzian compartment C (rectum) showed the highest rate (74%) of exact concordance,
i.e., the measurement by TVS matched exactly the evaluation during surgery. For #Enzian
compartment A (vagina, retrocervical area, and rectovaginal septum) and in particular
for #Enzian compartment B (uterosacral ligaments and parametria), the rates of exact
concordance were slightly lower (for example, 55% for B left and 60% for B right),
possibly because it may be difficult to differentiate fibrotic tissue surrounding
the DE lesion from the actual DE lesion during both TVS and surgery in some cases.
However, when a maximum tolerance margin of 3 mm was considered, a rate of 88% for
each side could be observed for compartment B, reaching the values seen for compartment
C.
Regarding #Enzian compartment O (ovary), the exact evaluation of the lesion size might
be particularly difficult during surgery. Endometriotic cysts are spherical and located
inside the ovary. It is not uncommon for the cyst to be opened when the ovary is mobilized,
and the next cyst is opened after aspiration of the contents. A direct measurement
of the dimensions is therefore more difficult. Furthermore, in the case of the presence
of more than one cyst in the same ovary, the maximal diameters of all cysts have to
be assessed to get the sum of the diameters, which determines the severity grade.
For these reasons, compared to #Enzian compartments A, B and C, the differences between
TVS and surgery are accordingly greater. However, greater differences may not be as
clinically relevant in the case of ovarian cysts as they are for endometriotic lesions
in other #Enzian compartments.
In #Enzian compartment T, subjective evaluation of the severity grade of the observed
adhesions might play a role to a certain extent during both TVS and surgery, in particular
for moderate adhesions (corresponding to #Enzian T2). However, in the presence of
severe adhesions (#Enzian T3) or slight adhesions (#Enzian T1), TVS and surgery showed
a high rate of concordance (86% of TVS T3 on the left, 84% of TVS T3 on the right,
as well as 75% of TVS T1 on the left and 77% of TVS T1 on the right side confirmed
by surgery). Therefore, severe adhesions seen on TVS are really severe or at least
moderate, and slight or absent adhesions on TVS do not turn out to be moderate or
severe, thus allowing for an adequate preoperative assessment of the anticipated surgical
complexity regarding tubo-ovarian adhesions.
When comparing the lesion sizes in compartments A, B, C, and O using the severity
grades of the #Enzian classification, similar results are obtained. The concordance
rates were slightly lower because discordant measurements during TVS and surgery often
differ by only a few millimeters but the two measurements belong to two adjacent severity
grades. For example, a 9 mm lesion on TVS will be classified as A1, but if it is 10
mm during surgery it will be classified as A2. This happens when measurements are
very close to 10 mm or 30 mm which are the limits between the different severity grades.
This can be seen most frequently for #Enzian compartments A and B, where lesions are
more often around 10 mm in size.
For #Enzian compartment FB (urinary bladder), all DE lesions seen on TVS were confirmed
by surgery, corresponding to a false-positive rate of 0% and a specificity of 100%.
Only one DE lesion was not detected by TVS, leading to a sensitivity of 86%. The only
missed lesion was an anterior nodule located between the urinary bladder and the anterior
abdominal wall measuring 10 mm in diameter. In #Enzian compartment FO there were only
two discordant findings: a lesion at the level of the umbilicus and another one in
the left broad ligament, which were seen during sonographic examination but could
not be confirmed by surgery. All other sonographic findings in #Enzian compartments
FB and FO were correct. In #Enzian compartments FI and FU, all sonographic findings
were confirmed by surgery (resulting in a sensitivity and specificity of 100% for
both compartments).
Despite the fact that the retrospective design might be considered a limitation of
this study, there were no missing data, and the comparison of different lesion sizes
could be carried out seamlessly. As a second limitation, the study uses the findings
of a tertiary referral center with an expert in gynecological sonography and experienced
surgeons. Therefore, the results may not be generally applicable to everyday clinical
practice. Thirdly, in the clinical setting of this study, the surgeon was not blinded
regarding the preoperative TVS evaluation. This may include a possible bias, as some
lesions in certain #Enzian compartments (mainly FI and FO) might only have been detected
during surgery because of the knowledge of their presence gained by preoperative sonographic
examination. On the other hand, this reflects real everyday clinical practice, and
it would have to be considered unethical and harmful for the patient to leave some
DE lesions because they were missed during surgery due to blinding of the surgeon
to the preoperative TVS.
In conclusion, the present study shows that preoperative evaluation of the location
and size of DE lesions in different #Enzian compartments by an expert in gynecological
sonography is very accurate, thereby providing the surgeon with a detailed depiction
of the extent of the disease that will be encountered during surgery. Furthermore,
by comparing the lesion sizes between TVS and surgery using the #Enzian severity grades,
it is the first study to show that the new #Enzian classification can be applied to
describe disease extent during both TVS and surgery, thereby offering a descriptive
system for both noninvasive and invasive specialties.
