diagnostic accuracy of transvaginal ultrasound for non-invasive diagnosis of bowel endometriosis

Ultrasound Obstet Gynecol 2011; 37: 257–263Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/uog.8858

Diagnostic accuracy of transvaginal ultrasound fornon-invasive diagnosis of bowel endometriosis: systematicreview and meta-analysis

G. HUDELIST*†, J. ENGLISH‡, A. E. THOMAS§, A. TINELLI¶, C. F. SINGER**and J. KECKSTEIN††*Department of Obstetrics and Gynaecology, Endometriosis and Pelvic Pain Clinic, Wilhelminen Hospital, Vienna, Austria; †SEF, StiftungEndometrioseforschung; ‡Department of Obstetrics and Gynaecology, University of Brighton Medical School, Brighton, UK; §Departmentof Methodological Research and Statistics, Institute of Psychology, Alpe Adria University Klagenfurt, Klagenfurt, Austria; ¶Department ofObstetrics and Gynaecology, Lecce Hospital, Lecce, Italy; **Department of Obstetrics and Gynaecology, University of Vienna, Vienna,Austria; ††Department of Obstetrics and Gynaecology, Center for Endometriosis, Villach Hospital, Villach, Austria

KEYWORDS: deep infiltrating endometriosis; presurgical diagnosis; transvaginal ultrasound

ABSTRACT

Objective To critically analyze the diagnostic valueof transvaginal sonography (TVS) for non-invasive,presurgical detection of bowel endometriosis.

Methods MEDLINE (1966–2010) and EMBASE (1980–2010) databases were searched for relevant studies investi-gating the diagnostic accuracy of TVS for diagnosing deepinfiltrating endometriosis involving the bowel. Diagnosiswas established by laparoscopy and/or histopathologi-cal analysis. Likelihood ratios (LRs) were recalculated inaddition to traditional measures of effectiveness.

Results Out of 188 papers, a total of 10 studies fulfilledpredefined inclusion criteria involving 1106 patientswith suspected endometriosis. The prevalence of bowelendometriosis varied from 24 to 73.3%. LR+ ranged from4.8 to 48.56 and LR− ranged from 0.02 to 0.36, with wideconfidence intervals. Pooled estimates of sensitivities andspecificities were 91 and 98%; LR+ and LR− were 30.36and 0.09; and positive and negative predictive values were98 and 95%, respectively. Three of the studies used bowelpreparations to enhance the visibility of the rectal wall;one study directly compared the use of water contrast vs.no prior bowel enema, for which the LR− was 0.04 and0.47, respectively.

Conclusions TVS with or without the use of priorbowel preparation is an accurate test for non-invasive,presurgical detection of deep infiltrating endometriosis ofthe rectosigmoid. Copyright 2011 ISUOG. Publishedby John Wiley & Sons, Ltd.

INTRODUCTION

Over the past decade, the use of transvaginal sono-graphy (TVS) has improved the quality of non-invasiveassessment of patients with suspected pelvic pathologies.With respect to endometriosis TVS has been shownto be a highly sensitive tool for the detection ofovarian endometriomas1 and is far superior to routineclinical examination alone2,3. Moore et al.1 systematicallyreviewed 67 papers on the validity of TVS for the detectionof pelvic endometriosis, out of which seven fulfilledthe inclusion criteria and focused on TVS imaging ofovarian endometriomas. The prevalence of the conditionranged between 13 and 38%. Sensitivities, specificities andpositive (LR+) and negative likelihood ratios (LR−) in sixstudies using gray-scale ultrasonography ranged between64 and 89%, 89 and 100%, 7.6 and 29.8 and 0.1 and 0.4,respectively. The authors therefore concluded that TVSshould be regarded as a useful test for identifying cysticovarian endometriosis presurgically.

Recent studies also suggest that TVS could be anaccurate method for the detection of endometriosis inextra-ovarian locations, i.e. uterosacral ligament involve-ment, endometriosis of the rectovaginal space, the pouchof Douglas, the vagina, the urinary bladder and deep infil-trating endometriosis (DIE) of the rectosigmoid3–8. SinceTVS is a readily available, cost- and time-effective diag-nostic instrument when compared to other radiologicalprocedures such as computed tomography and magneticresonance imaging (MRI)9,10, several investigators havefurther examined the diagnostic value of TVS for the non-invasive detection of DIE infiltrating the bowel. The aim

Correspondence to: Prof. G. Hudelist, Department of Obstetrics and Gynaecology, Endometriosis and Pelvic Pain Clinic, WilhelminenGeneral Hospital, Montlearstrasse 37, A-1160 Vienna, Austria (e-mail: gernot [email protected])

Accepted: 7 October 2010

Copyright 2011 ISUOG. Published by John Wiley & Sons, Ltd. SYSTEMATIC REVIEW

258 Hudelist et al.

of this work was to systematically analyze the publishedliterature evaluating the role of TVS for the detection ofDIE involving the rectosigmoid.

METHODS

The MEDLINE (1966–2010) and EMBASE (1980–2010)databases were searched using the following searchstrategy:

1. (pelvic or ovarian or deep infiltrating) near2 (mass orcyst* or tumo*r)

2. ‘ENDOMETRIOSIS’ in MeSH or 1 or ‘BOWELENDOMETRIOSIS’

3. 2, not case reports, not review articles4. with checktags ‘female’ and ‘human’5. with ‘ULTRASOUND’/all subheadings or ‘TRANS-

VAGINAL’/all subheadings or ‘SONOGRAPHY’

Abstracts of all studies identified were read andmanuscripts were then fully reviewed. In addition,reference lists of all reviewed manuscripts were searchedfor additional data. Study selection and assessment ofquality were performed independently by two reviewers(G. H. and J. E).

Selection criteria

All studies included in the present review had to beprospective and were required to involve both TVS exam-ination and surgical exploration of the pelvis eitherby laparoscopy or by laparotomy (as stated by Mooreet al.1). Scientific publications including case reports, stud-ies on adenomyosis or extrapelvic endometriotic diseaseas well as retrospective case series and review articleswere excluded. Studies reporting on pregnant women,rectal ultrasound as the only examination and endoscopicsonography were also excluded from this review. Patientsincluded in the studies presented with either subfertilityor symptoms suggestive of endometriosis.

According to the criteria of Moore et al.1, studies wereconsidered to be of good quality when information onrecruitment of patients, blinding of ultrasound operatorsand surgeons and data on the technical equipment wereprovided. In order to define the stage and severity ofdisease (i.e. the final endpoint of diagnosis), studies hadto describe the anatomical location of deep infiltratingdisease combined with histological confirmation ofendometriosis. Moore et al.1 considered that studiesmissing one or two of these criteria were of moderate/poorquality.

Data extraction and presentation of results

As described by Moore et al.1, 2 × 2 tables were created tovalidate test results against surgical and histopathologicalfindings aimed at defining whether DIE involving therectosigmoid can be detected by TVS. In addition,QUADAS (quality assessment of diagnostic accuracy of

studies) was used to assess the studies11. As describedpreviously, study quality is defined as high when ≥ 9items out of 14 are met, moderate when 6 items aremet and low when < 6 criteria are met12. LR+, LR−and test accuracy were calculated in studies lackingthese data. Confidence intervals (CIs) were calculatedas described previously1 using CATMaker statisticalsoftware (Centre for Evidence-Based Medicine, Oxford,UK). In two cases raw data (true and false positiveand negative rates) were obtained from the authors.Since heterogeneity is a common finding in diagnosticmeta-analyses we calculated Cochran’s Q and I2 forall measures to assess the significance and magnitudeof study heterogeneity13. A forest plot was used toassess eventual outlier studies (data not shown). Potentialsources of heterogeneity were explored by random-effectsmeta-regression. Number of subjects, year of publication,QUADAS scores and prevalence were used as study-levelcovariates that predicted the logarithm of diagnostic oddsratios (DORs). The studies were weighted by the inverse ofvariance of the DOR to consider the precision with whicheach study measured it. Relative DORs were calculated tocompare the overall DOR with the adjusted DOR. Whenheterogeneity between studies is present, a random-effectsmodel can be used to obtain pooled estimates14.

We applied random-effects models with the DerSir-monian–Laird estimator in order to determine overallestimates of sensitivity, specificity, LR+, LR− and DOR.For data with zero counts a continuity correction of 0.5was added to every value in that study, thereby allowingthe calculation of all LRs15. For the assessment of pub-lication bias Egger’s test and Begg’s test were conductedand funnel plots were investigated (data not shown). Thetests for publication bias and funnel plots were performedwith R statistical software version 2.11.1 metafor pack-age (R Development Core Team, Vienna, Austria)16. Allother analyses were performed with MetaDiSc statisticalsoftware version 1.417 (Hospital Universitario, Madrid,Spain).

RESULTS

The initial implementation of the research strategyrevealed 188 studies relating to endometriosis and/oradenomyosis and/or ovarian endometriosis diagnosedby laparoscopy or laparotomy and/or ultrasonography.Out of these only 51 papers specifically used TVSand surgical exploration to diagnose DIE. Of these 51papers, seven were excluded because they were casereports or descriptive in nature. A further 18 papersdid not meet the inclusion criteria due to the factthat they were review articles, despite the exclusionof this article type in the primary search process.Finally, three other publications included comments onpublications and were also excluded from the finalanalysis, leaving 23 manuscripts for review3–5,7,8,18–35.Out of these 23 papers, 13 publications were excludeddue to methodological problems; three papers werepurely retrospective in nature18,27,28; four manuscripts

Copyright 2011 ISUOG. Published by John Wiley & Sons, Ltd. Ultrasound Obstet Gynecol 2011; 37: 257–263.

Transvaginal ultrasound in the diagnosis of bowel endometriosis 259

were excluded because they involved a group ofpatients already described in a previous publicationby the same group of authors5,7,25,29 and anotherfour papers were not included in the analysis becauseinsufficient information was provided on the presenceor absence of bowel endometriosis within the group ofwomen with endometriosis22,30,32,34. Two papers wereexcluded due to the low number of cases with provenbowel endometriosis24 or insufficient information on theanatomical localization of DIE affecting the bowel35. Thus10 papers fulfilled the inclusion criteria and were includedin the final review. In addition to the quality criteriadescribed by Moore et al.1, the QUADAS scores werecalculated for each of the studies. These ranged from 7 to13, reflecting high methodological quality in eight casesand moderate quality in two cases.

Out of these, four papers evaluated the diagnosticvalue of TVS for the detection of DIE infiltrating therectosigmoid excluding other possible locations of DIE inthe analysis8,23,31,33. All other works also evaluated thepotential of TVS for the visualization of other affectedsites of deep infiltrating disease such as the vagina,rectovaginal space, uterosacral ligaments, bladder andpouch of Douglas. Two papers addressed the use ofTVS to diagnose DIE involving the rectum, pouch ofDouglas, vagina, rectovaginal space, uterosacral ligamentsand bladder3,26 using laparoscopic exploration andhistological confirmation of endometriosis as the goldstandard test. One work compared TVS with MRI todetect DIE preoperatively compared to the gold standardlaparoscopy without histological confirmation21. Onepaper assessed the diagnostic value of TVS vs. rectalendosonography19; two papers compared TVS with MRI,digital examination4 and rectal endosonography20 todetect DIE presurgically. Data on the quality of thesestudies including number of patients and cases of bowelendometriosis, recruitment criteria, blinding, informationand presence of diagnostic criteria and technical details,data on reference standards and assessment of thegrade of study quality are provided in Table 1. Table 2depicts prevalence rates, sensitivities, specificities, testaccuracies, positive predictive values (PPVs), negativepredictive values (NPVs) and positive and negative LRswith CIs of all studies included in the final analysis. Thenumber of patients with proven DIE infiltrating the bowelundergoing preoperative TVS ranged from 1721 to 8123.Sensitivity and specificity varied from 67 to 98% and 92to 100%, respectively.

As further demonstrated in Table 2, PPVs and NPVsranged from 87 to 100% and 75 to 99%, respectively.LR+ goes up to infinity in cases lacking false-positivefindings and thus are not shown. LR+ of the remainingstudies varied from 4.8 to 48.56 and LR− from 0.02 to0.36.

Three studies used TVS combined with a bowel enemato provide a better visualization of rectal wall anatomy,specifically muscularis propria4,23,33. Valenzano Menadaet al.33 directly compared whether adding water-contrastin the rectosigmoid (RWC-TVS) during TVS improves

presurgical diagnosis of rectal DIE with TVS. Thesensitivity of RWC-TVS vs. TVS in identifying rectal DIEwas 97 vs. 56%, the specificity 100 vs. 92.5%, the PPV100 vs. 72% and the NPV 99 vs. 86%. Due to the absenceof false-positive cases in the RWC-TVS cohort, LR+ couldnot be calculated, and LR− was 0.04 vs. 0.47 for TVS.

Meta-analysis of all studies included in the final reviewyielded significant results of Cochran’s Q for all measuresexcept LR+ (sensitivity: P < 0.001, Q = 238.64, df =9, I2 = 76.7%; specificity: P = 0.037, Q = 17.8, df =9, I2 = 49.5%; LR+: P = 0.144, Q = 13.4, df = 9,I2 = 33.0%; LR−: P < 0.001, Q = 52.6, df = 9, I2 =82.9%; DOR: P = 0.001, Q = 27.6, df = 9, I2 = 67.4%),indicating considerable heterogeneity between studies.Significant values for study heterogeneity were mainlycaused by two studies21,26 that were aberrant in respectto sensitivity, LR− and DOR. However, meta-regressionincluding sample size, prevalence, QUADAS and year ofpublication did not yield any significant results: numberof subjects (P = 0.197, relative DOR = 1.02 (95% CI,0.99–1.04)), prevalence (P = 0.568, relative DOR =0.97 (95% CI, 0.88–1.08)), QUADAS score (P = 0.319,relative DOR = 1.44 (95% CI, 0.65–3.20)) and yearof publication (P = 0.298, relative DOR = 1.35 (95%CI, 0.67–2.71)). This suggests that although there wasrelevant heterogeneity between the studies, the influenceof the covariates was not systematic. To account forheterogeneity we used a random-effects model to performpooled estimates and estimate the respective CIs (Table 3).As all studies were conducted with women who sufferedfrom pain or infertility, the pooled prevalence of 47%refers to women with specific symptoms.

Evaluation by both Egger’s test (P = 0.221) and Begg’stest (P = 0.293) did not show evidence of publicationbias for logDOR. This result was confirmed by inspectionof the funnel plots, which were all symmetrical for theinvestigated diagnostic measures (sensitivity, specificity,PPV, NPV, LR+, LR− and accuracy; data not shown).

DISCUSSION

Endometriosis infiltrating the rectosigmoid can besuspected in up to 9–22% of all women with provenendometriosis36,37. Symptoms of DIE involving the bowelvary greatly, ranging from asymptomatic women withextensive rectal involvement to patients with severedysmenorrhea and dyschezia38. Treatment strategiesinclude hormonal preparations or surgical excision ofendometriotic nodules37,39, but presurgical staging ofDIE is crucial for planning surgical treatment options.The findings of our systematic review clearly suggestthat TVS is a highly valuable tool for the non-invasivedetection of DIE affecting the rectosigmoid. In addition tosensitivities, specificities, PPVs, NPVs and test accuracies,we recalculated all positive and negative LRs since thesereflect the diagnostic accuracy and the clinical usefulnessof a test independently of the prevalence of the studycondition in the study population1. The prevalence of


260 Hudelist et al.

Tab

le1

Cha

ract

eris

tics

ofth

est

udie

sin

clud

edin

the

anal

ysis

Stud

y

Dia

gnos

tic

crit

eria

stat

ed;

ultr

asou

ndcr

iter

iafo

rdi

agno

sis

Stud

yde

sign

Blin

ding

Rec

ruit

men

t;ca

use

for

refe

rral

Pat

ient

sw

ith

susp

ecte

den

dom

etri

osis

(n)

Cas

esof

rect

al/s

igm

oida

len

dom

etri

osis

(n)

Tes

tm

etho

dR

efer

ence

stan

dard

Stud

yqu

alit

yac

cord

ing

toM

oore

crit

eria

(QU

AD

AS

scor

e)

Baz

otet

al.

(200

3)19

Yes

;thi

cken

ing

ofth

em

uscu

lari

spr

opri

a>

3m

m,

whi

chis

hypo

echo

ican

dth

in(<

3m

m)

unde

rno

rmal

circ

umst

ance

s

Pros

pect

ive

Sono

grap

hers

blin

ded

Con

secu

tive

;pai

nan

din

fert

ility

3022

TV

S,R

ES

His

tolo

gyM

oder

ate

(9)

Baz

otet

al.

(200

4)3

Yes

;as

abov

ePr

ospe

ctiv

eSo

nogr

aphe

rbl

inde

dC

onse

cuti

ve;p

ain

and

infe

rtili

ty14

247

TV

SH

isto

logy

in29

Mod

erat

e(8

)

Car

bogn

inet

al.

(200

6)21

Not

desc

ribe

dPr

ospe

ctiv

eN

otst

ated

Con

secu

tive

;pai

nan

din

fert

ility

3217

TV

S,M

RI

Lap

aros

copi

cvi

sual

izat

ion

Poor

(7)

Abr

aoet

al.

(200

7)4

Yes

;nod

ular

,pre

dom

inan

tly

solid

,hyp

oech

ogen

icle

sion

adhe

red

toth

ew

allo

fth

ein

test

inal

loop

Pros

pect

ive

Sono

grap

her

blin

ded

Con

secu

tive

;pai

nan

din

fert

ility

104

54T

VS,

MR

I,PV

His

tolo

gyM

oder

ate

(10)

Val

enza

noM

enad

aet

al.

(200

8)33

Yes

;rec

tova

gina

lhyp

oech

oic

mas

sad

here

ntan

d/or

pene

trat

edin

toth

ein

test

inal

wal

lthi

cken

ing

the

mus

cula

ris

muc

osa

Pros

pect

ive

Sono

grap

hers

blin

ded

Con

secu

tive

;pai

nan

din

fert

ility

9023

TV

S, RW

C-T

VS

His

tolo

gyG

ood

(12)

Gue

rrie

roet

al.

(200

8)26

Yes

;pre

senc

eof

som

eth

inba

nd-l

ike

echo

esde

part

ing

from

the

cent

erof

the

mas

sas

an‘I

ndia

nhe

addr

ess’

.

Pros

pect

ive

No

info

rmat

ion

give

nC

onse

cuti

ve;p

ain

and

infe

rtili

ty88

39T

VS

His

tolo

gyG

ood

(11)

Pike

tty

etal

.(2

009)

31Y

es;i

rreg

ular

hypo

echo

icm

ass,

wit

hor

wit

hout

hypo

/hyp

erec

hoic

foci

invo

lvin

gco

lon

mus

cula

ris

Pros

pect

ive

Sono

grap

her

blin

ded

Con

secu

tive

;pai

nan

din

fert

ility

134

75T

VS

His

tolo

gyG

ood

(12)

Baz

otet

al.

(200

9)20

Yes

;as

abov

ePr

ospe

ctiv

eR

adio

logi

sts

blin

ded

Con

secu

tive

;pai

nan

din

fert

ility

9263

TV

S,M

RI,

PV,R

ES

His

tolo

gyin

54G

ood

(12)

Hud

elis

tet

al.

(200

9)8

Yes

;as

abov

ePr

ospe

ctiv

eSo

nogr

aphe

rno

tbl

inde

dto

PVC

onse

cuti

ve;p

ain

and

infe

rtili

ty20

048

TV

S,PV

His

tolo

gyM

oder

ate

(13)

Gon

calv

eset

al.

(201

0)23

Yes

;lon

g,no

dula

r,pr

edom

inan

tly

solid

,hy

poec

hoge

nic

lesi

onad

here

dto

the

wal

lof

the

inte

stin

allo

op

Pros

pect

ive

Sono

grap

her

blin

ded

Con

secu

tive

;pai

nan

din

fert

ility

194

81T

VS

His

tolo

gyG

ood

(12)

MR

I,m

agne

tic

reso

nanc

eim

agin

g;PV

,per

vagi

nam

clin

ical

exam

inat

ion;

QU

AD

AS,

qual

ity

asse

ssm

ent

ofdi

agno

stic

accu

racy

ofst

udie

s;R

ES,

rect

alen

doso

nogr

aphy

;RW

C,r

ecta

lwat

erco

ntra

st;

TV

S,tr

ansv

agin

also

nogr

aphy

.



Tab

le2

Ana

lysi

sof

data

inth

ein

clud

edst

udie

s

Stud

y

Pre

vale

nce

ofre

ctal

/sig

moi

dal

endo

met

rios

is(n

(%))

Sens

itiv

ity

(n(%

))Sp

ecifi

city

(n(%

))P

PV

(%)

NP

V(%

)A

ccur

acy

(%)

LR

+(9

5%C

I)L

R−

(95%

CI)

Baz

otet

al.(

2003

)1922

/30

(73)

21/2

2(9

5)8/

8(1

00)

100

8997

—0.

05(0

.01

–0.3

1)B

azot

etal

.(20

04)3

47/1

42(3

3)41

/47

(87)

92/9

5(9

7)93

9494

27.6

2(9

.02

–84.

58)

0.13

(0.0

6–0

.28)

Car

bogn

inet

al.(

2006

)2117

/32

(53)

12/1

7(7

1)15

/15

(100

)10

075

844.

8(1

.26

–18.

31)

0.29

(0.1

4–0

.61)

Abr

aoet

al.(

2007

)454

/104

(52)

53/5

4(9

8)50

/50

(100

)10

098

99—

0.02

(0.0

0–0

.13)

Val

enza

noM

enad

aet

al.(

2008

)3323

/90

(26)

22/2

3(9

6)67

/67

(100

)10

099

99—

0.04

(0.0

1–0

.3)

Gue

rrie

roet

al.(

2008

)2639

/88

(44)

26/3

9(6

7)45

/49

(92)

8778

818.

17(3

.11

–21.

44)

0.36

(0.2

3–0

.57)

Pike

tty

etal

.(20

09)31

75/1

33(5

6)68

/75

(91)

56/5

8(9

7)97

8993

26.2

9(6

.72

–102

.83)

0.10

(0.0

5–0

.20)

Baz

otet

al.(

2009

)2063

/92

(68)

59/6

3(9

4)29

/29

(100

)10

088

96—

0.06

(0.0

2–0

.16)

Hud

elis

tet

al.(

2009

)848

/200

(24)

46/4

8(9

6)14

9/15

2(9

8)94

9998

48.5

6(1

5.81

–149

.10)

0.04

(0.0

1–0

.17)

Gon

calv

eset

al.(

2010

)2381

/194

(42)

79/8

1(9

8)11

3/11

3(1

00)

100

9899

—0.

02(0

.01

–0.1

0)

LR

+,po

siti

velik

elih

ood

rati

o;L

R−,

nega

tive

likel

ihoo

dra

tio;

NPV

,neg

ativ

epr

edic

tive

valu

e;PP

V,p

osit

ive

pred

icti

veva

lue.

Table 3 Overall analysis of all studies included in the final analysisusing a random-effects model to perform pooled estimates ofvariables

Variable Estimate (95% CI)

Sensitivity (%) 91 (88.1–93.5)Specificity (%) 98 (96.7–99.0)LR+ 30.36 (15.457–59.626)LR− 0.09 (0.046–0.188)DOR 394.3 (116.3–1336.0)Prevalence (%) 47 (36.7–57.3)PPV (%) 98 (96.7–99.6)NPV (%) 95 (92.1–97.7)

DOR, diagnostic odds ratio (values not adjusted); LR+, positivelikelihood ratio (with continuity correction for studies withnull-cells); LR−, negative likelihood ratio; NPV, negative predictivevalue; PPV, positive predictive value.

DIE involving the bowel ranged from 24 to 73%, whichmay be attributable to different referral patterns and theavailability of tertiary referral centers providing sufficientexpertise in the presurgical diagnosis and treatment ofrectal DIE. LR+ ranged from 4.821 to 27.623 and LR−from 0.0223 to 0.3626. The combined use of TVS withvaginal examination may further increase the diagnosticaccuracy, with reported LR+ and LR− 48.56 and 0.048,respectively. In addition, it should be noted that LR+went up to infinity in five studies (Table 2), suggestinghighly accurate presurgical test results.

Performance of a meta-analysis of all included studiesby using a random-effects model to calculate pooledestimates revealed LR+ and LR− of 30.36 and 0.09,respectively. According to Altman40, a very helpful testfor establishing or excluding a condition is characterizedby an LR+ > 10 and an LR− < 0.1, while a test isregarded as moderately helpful with an LR+ between 5and 10 and an LR− between 0.1 and 0.2. As depicted inTable 3, the results of our meta-analysis clearly suggestthat TVS is indeed very useful for sonographic diagnosisbut also presurgical exclusion of bowel endometriosis.Two studies21,26 were aberrant with respect to sensitivity,LR− and DOR. This may be attributable to low sensitivityvalues in association with small patient numbers21 orthe use of ‘tenderness guided’ transvaginal sonography26,which may be less sensitive in cases of endometrioticinvolvement of the upper rectum/lower sigmoid, sincethese locations may not appear as painful sites whenusing this technique.

The conclusions of this review are weakened by thefact that blinding of the surgeon was missing in allstudies included in the final analysis, which might havea potential influence on the results of the gold standardtest, i.e. surgery. On the other hand, information onthe preoperative findings is essential for guiding thesurgical technique and surgical diagnosis of endometriosisin everyday clinical practice.

An additional variable in the accuracy of the goldstandard test, i.e. laparoscopy, is the fact that severalstudies did not provide sufficient information about the


262 Hudelist et al.

surgical exploration of the patients with DIE3,20,21,26

since complete cul-de-sac obliteration secondary toendometriosis was reported but could not necessarilybe surgically cleared. This may, in cases of DIEaffecting the non-visible, occluded lower rectum, leadto misinterpretation and observation bias of the goldstandard test and consequently negatively affect thevalidation of the index test, i.e. TVS. Hence, full surgicalexploration of patients with an occluded pouch of Douglasis important for the final and accurate diagnosis ofbowel endometriosis since DIE may be missed at thetime of laparoscopy, thereby questioning the qualityof laparoscopic visualization of the pelvis as the goldstandard test for the diagnosis of endometriosis.

Another limiting factor to the results of this systematicreview is the heterogeneity of populations included inthe studies. Some authors failed to provide sufficientdata on the selection criteria and referral patterns ofcenters where studies were conducted. Due to the factthat most patients were treated in tertiary referral centers,it is conceivable that the study populations represent analready preselected patient cohort with a high prevalenceof bowel endometriosis. As a consequence, the conclusionsof this review may therefore only be applicable to a tertiaryreferral setting and patients with a high risk for DIE.

In clinical everyday practice, exclusion of DIE is ofcritical importance since extensive bowel involvementwarrants an interdisciplinary approach and referral toa tertiary center. In conclusion, the majority of thepublished evidence suggests that TVS is a highly usefuland easy accessible test for the preoperative detection ofDIE infiltrating the rectosigmoid. Whether the widespreaduse of TVS and the inclusion of TVS in specialist trainingprograms lead to a reduction in the diagnostic delay ofpatients with endometriosis remains to be seen.

ACKNOWLEDGMENTS

The authors want to thank Dr Simone Ferrero and MMagNadja Fritzer for their support and helpful advice. Thiswork was supported by the OEGEO, OsterreichischeGesellschaft fur Endokrinologische Onkologie.

REFERENCES

1. Moore J, Copley S, Morris J, Lindsell D, Golding S, Kennedy S.A systematic review of the accuracy of ultrasound in thediagnosis of endometriosis. Ultrasound Obstet Gynecol 2002;20: 630–634.

2. Nezhat C, Santolaya J, Nezhat FR. Comparison of transvaginalsonography and bimanual pelvic examination in patients withlaparoscopically confirmed endometriosis. J Am Assoc GynecolLaparosc 1994; 1: 127–130.

3. Bazot M, Thomassin I, Hourani R, Cortez A, Darai E. Diag-nostic accuracy of transvaginal sonography for deep pelvicendometriosis. Ultrasound Obstet Gynecol 2004; 24: 180–185.

4. Abrao MS, Goncalves MO, Dias JA Jr, Podgaec S, Chamie LP,Blasbalg R. Comparison between clinical examination, trans-vaginal sonography and magnetic resonance imaging for thediagnosis of deep endometriosis. Hum Reprod 2007; 22:3092–3097.

5. Bazot M, Malzy P, Cortez A, Roseau G, Amouyal P, Darai E.Accuracy of transvaginal sonography and rectal endoscopicsonography in the diagnosis of deep infiltrating endometriosis.Ultrasound Obstet Gynecol 2007; 30: 994–1001.

6. Hudelist G, Keckstein J. The use of transvaginal sonography(TVS) for preoperative diagnosis of pelvic endometriosis. Praxis(Bern 1994) 2009; 98: 603–607.

7. Hudelist G, Oberwinkler KH, Singer CF, Tuttlies F, Rauter G,Ritter O, Keckstein J. Combination of transvaginal sonographyand clinical examination for preoperative diagnosis of pelvicendometriosis. Hum Reprod 2009; 24: 1018–1024.

8. Hudelist G, Tuttlies F, Rauter G, Pucher S, Keckstein J. Cantransvaginal sonography predict infiltration depth in patientswith deep infiltrating endometriosis of the rectum? Hum Reprod2009; 24: 1012–1017.

9. Kinkel K, Frei KA, Balleyguier C, Chapron C. Diagnosis ofendometriosis with imaging: a review. Eur Radiol 2006; 16:285–298.

10. Brosens I, Puttemans P, Campo R, Gordts S, Kinkel K. Diagno-sis of endometriosis: pelvic endoscopy and imaging techniques.Best Pract Res Clin Obstet Gynaecol 2004; 18: 285–303.

11. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. Thedevelopment of QUADAS: a tool for the quality assessment ofstudies of diagnostic accuracy included in systematic reviews.BMC Med Res Methodol 2003; 3: 25.

12. Meredith SM, Sanchez-Ramos L, Kaunitz AM. Diagnosticaccuracy of transvaginal sonography for the diagnosis of ade-nomyosis: systematic review and metaanalysis. Am J ObstetGynecol 2009; 201: 107.e1–6.

13. Sidik K, Jonkman JN. Simple heterogeneity variance estimationin meta-analysis. J R Stat Soc [Ser C] 2005; 54: 367–384.

14. Deville W, Buntix F, Bouter LM, Montori VM, de Vet HC, vander Windt DA, Bezemer PD. Conductiong systematic reviews ofdiagnostic studies: didactic guidelines. BMC Med Res Methodol2002; 2: 2–9.

15. Goodacre S, Sampson F, Thomas S, van Beek E, Sutton A.Systematic review and meta-analysis of the diagnostic accuracyof ultrasonography for deep vein thrombosis. BMC MedImaging 2005; 5: 5–6.

16. Viechtbauer W. Conducting meta-analysis in R with themetaphor package. J Stat Softw 2010; 36: 1–48.

17. Zamora J, Abraira V, Muriel A, Khan K. Coomarasamy Meta-DiSc: a software for meta-analysis of test accuracy data. BMCMed Res Methodol 2006; 6: 6–12.

18. Balleyguier C, Chapron C, Dubuisson JB, Kinkel K, Faucon-nier A, Vieira M, Helenon O, Menu Y. Comparison of magneticresonance imaging and transvaginal ultrasonography in diag-nosing bladder endometriosis. J Am Assoc Gynecol Laparosc2002; 9: 15–23.

19. Bazot M, Detchev R, Cortez A, Amouyal P, Uzan S, Darai E.Transvaginal sonography and rectal endoscopic sonographyfor the assessment of pelvic endometriosis: a preliminarycomparison. Hum Reprod 2003; 18: 1686–1692.

20. Bazot M, Lafont C, Rouzier R, Roseau G, Thomassin-NaggaraI, Darai E. Diagnostic accuracy of physical examination,transvaginal sonography, rectal endoscopic sonography, andmagnetic resonance imaging to diagnose deep infiltratingendometriosis. Fertil Steril 2009; 92: 1825–1833.

21. Carbognin G, Girardi V, Pinali L, Raffaelli R, Bergamini V,Pozzi Mucelli R. Assessment of pelvic endometriosis: correlationof US and MRI with laparoscopic findings. Radiol Med 2006;111: 687–701.

22. Eskenazi B, Warner M, Bonsignore L, Olive D, Samuels S,Vercellini P. Validation study of nonsurgical diagnosis ofendometriosis. Fertil Steril 2001; 76: 929–935.

23. Goncalves MO, Podgaec S, Dias JA Jr, Gonzalez M, Abrao MS.Transvaginal ultrasonography with bowel preparation is able topredict the number of lesions and rectosigmoid layers affectedin cases of deep endometriosis, defining surgical strategy. HumReprod 2010; 25: 665–671.



24. Grasso RF, Di Giacomo V, Sedati P, Sizzi O, Florio G, Faiella E,Rossetti A, Del Vescovo R, Zobel BB. Diagnosis of deepinfiltrating endometriosis: accuracy of magnetic resonanceimaging and transvaginal 3D ultrasonography. Abdom Imaging2009; 35: 716–725.

25. Guerriero S, Ajossa S, Gerada M, D’Aquila M, Piras B, MelisGB. Tenderness-guided transvaginal ultrasonography: a newmethod for the detection of deep endometriosis in patients withchronic pelvic pain. Fertil Steril 2007; 88: 1293–1297.

26. Guerriero S, Ajossa S, Gerada M, Virgilio B, AngioniS, Melis GB. Diagnostic value of transvaginal ‘tenderness-guided’ ultrasonography for the prediction of location of deependometriosis. Hum Reprod 2008; 23: 2452–2457.

27. Hensen JH, Puylaert JB. Endometriosis of the posterior cul-de-sac: clinical presentation and findings at transvaginalultrasound. AJR Am J Roentgenol 2009; 192: 1618–1624.

28. Koga K, Osuga Y, Yano T, Momoeda M, Yoshino O, Hirota Y,Kugu K, Nishii O, Tsutsumi O, Taketani Y. Characteristicimages of deeply infiltrating rectosigmoid endometriosis ontransvaginal and transrectal ultrasonography. Hum Reprod2003; 18: 1328–1333.

29. Menada MV, Remorgida V, Abbamonte LH, Fulcheri E, RagniN, Ferrero S. Transvaginal ultrasonography combined withwater-contrast in the rectum in the diagnosis of rectovaginalendometriosis infiltrating the bowel. Fertil Steril 2008; 89:699–700.

30. Okaro E, Condous G, Khalid A, Timmerman D, Ameye L,Huffel SV, Bourne T. The use of ultrasound-based ‘soft markers’for the prediction of pelvic pathology in women with chronicpelvic pain – can we reduce the need for laparoscopy? BJOG2006; 113: 251–256.

31. Piketty M, Chopin N, Dousset B, Millischer-Bellaische AE,Roseau G, Leconte M, Borghese B, Chapron C. Preoperativework-up for patients with deeply infiltrating endometriosis:transvaginal ultrasonography must definitely be the first-lineimaging examination. Hum Reprod 2009; 24: 602–607.

32. Savelli L, Manuzzi L, Pollastri P, Mabrouk M, Seracchioli R,Venturoli S. Diagnostic accuracy and potential limitations oftransvaginal sonography for bladder endometriosis. UltrasoundObstet Gynecol 2009; 34: 595–600.

33. Valenzano Menada M, Remorgida V, Abbamonte LH, Nico-letti A, Ragni N, Ferrero S. Does transvaginal ultrasonographycombined with water-contrast in the rectum aid in the diagno-sis of rectovaginal endometriosis infiltrating the bowel? HumReprod 2008; 23: 1069–1075.

34. Yazbek J, Helmy S, Ben-Nagi J, Holland T, Sawyer E, JurkovicD. Value of preoperative ultrasound examination in the selectionof women with adnexal masses for laparoscopic surgery.Ultrasound Obstet Gynecol 2007; 30: 883–888.

35. Holland TK, Yazbek J, Cutner A, Saridogan E, Hoo WL,Jurkovic D. Value of transvaginal ultrasound in assessing theseverity of pelvic endometriosis. Ultrasound Obstet Gynecol2010; 36: 241–248.

36. Chapron C, Chopin N, Borghese B, Foulot H, Dousset B,Vacher-Lavenu MC, Vieira M, Hasan W, Bricou A. Deeplyinfiltrating endometriosis: pathogenetic implications of theanatomical distribution. Hum Reprod 2006; 21: 1839–1845.

37. Chapron C, Fauconnier A, Vieira M, Barakat H, Dousset B,Pansini V, Vacher-Lavenu MC, Dubuisson JB. Anatomical dis-tribution of deeply infiltrating endometriosis: surgical implica-tions and proposition for a classification. Hum Reprod 2003;18: 157–161.

38. Ballard K, Lane H, Hudelist G, Banerjee S, Wright J. Canspecific pain symptoms help in the diagnosis of endometriosis?A cohort study of women with chronic pelvic pain. Fertil Steril2010; 94: 20–27.

39. Keckstein J, Ulrich U, Kandolf O, Wiesinger H, Wustlich M.Laparoscopic therapy of intestinal endometriosis and theranking of drug treatment. Zentralbl Gynakol 2003; 125:259–266.

40. Altman DG. Statistics in medical journals: developments in the1980s. Stat Med 1991; 10: 1897–1913.


diagnostic accuracy of transvaginal ultrasound for non-invasive diagnosis of bowel endometriosis

Documents