SYSTEMATIC REVIEW AND META-ANALYSIS

Diagnostic accuracy of EUS compared with MRCP in detecting

choledocholithiasis: a meta-analysis of diagnostic test accuracy
in head-to-head studies
Yaser Meeralam, MD, Khalil Al-Shammari, MD, Mohammad Yaghoobi, MD, MSc (Epi), AFS, DABIM, FRCPC
Hamilton Ontario, Canada

Background and Aims: There is a wide range of reported sensitivity and specificity for EUS and MRCP in
the diagnosis of choledocholithiasis, with lack of a proper meta-analysis of diagnostic test accuracy by
using head-to-head comparison. Here, we aimed to compare the diagnostic accuracy of EUS and MRCP in
detecting choledocholithiasis by using appropriate methodology recommended by the Cochrane
Collaboration.
Methods: A comprehensive electronic literature search up to January 2017 was done by 2 reviewers for
prospective cohort studies comparing EUS and MRCP to a reference standard for detecting choledocholithiasis.
The acceptable reference standards were considered ERCP, intraoperative cholangiography, or clinical follow-up
>3 months for negative cases. Quality of the included studies was measured by using the QUADAS-2 tool. A bivar-
iate hierarchical model was used to perform the meta-analysis of diagnostic test accuracy. Summary receiver oper-
ating characteristics were developed and the area under the curve was calculated.
Results: A total of 5 of 32 studies were included. No study presented a high risk of bias. The pooled sensitivity
and specificity were 0.97 (range, 0.91-0.99) and 0.90 (range, 0.83-0.94) for EUS and 0.87 (range, 0.80-0.93) and
0.92 (range, 0.87-0.96) for MRCP. The overall diagnostic odds ratio of EUS was significantly higher than the
one with MRCP (162.5 vs 79.0, respectively; P Z .008). Further analysis showed that this was mainly due to
the significantly higher sensitivity of EUS as compared with that of MRCP (P Z .006). The specificity was not
significantly different between 2 modalities (P Z .42).
Conclusion: Both EUS and MRCP provide good diagnostic accuracy, with EUS providing statically better
diagnostic accuracy and sensitivity, with comparable specificity. EUS should be incorporated in the
diagnostic algorithm in patients suspected of choledocholithiasis whenever appropriate. (Gastrointest En-
dosc 2017;86:986-93.)

Choledocholithiasis is present in 6% to 15% of the initial investigations for suspected choledocholithiasis;

general population and can cause abdominal pain and
jaundice or end in potentially life-threatening outcomes
such as acute cholangitis or acute pancreatitis.1-4
Clinical assessment and abdominal US often are the
Abbreviations: AUC, area under the curve; DOR, diagnostic odds ratio;
DTA, diagnostic test accuracy; IOC, intraoperative cholangiography; LR,
likelihood ratio; sROC, summary receiver operating characteristic curve.
DISCLOSURE: All authors disclosed no financial relationships relevant
to this publication.
See CME section; p. 1151.
Copyright ª 2017 by the American Society for Gastrointestinal Endoscopy
0016-5107/$36.00
http://dx.doi.org/10.1016/j.gie.2017.06.009
(footnotes continued on last page of article)
however, they lack sensitivity and specificity. EUS and
MRCP are more accurate in the diagnosis of choledocho-
lithiasis.5 ERCP and intraoperative cholangiography
(IOC) usually are considered the reference standards.6-
8
However, they are associated with significant adverse
events such as post-ERCP pancreatitis, and their routine
use in the diagnosis of choledocholithiasis is not recom-
mended.9,10 EUS and MRCP are known to have lower
adverse event rates and to provide acceptable sensitivity
and specificity for choledocholithiasis.5 A meta-analysis
of 27 studies showed a pooled sensitivity of 0.94 (95%
confidence interval [CI], 0.93-0.96) and a specificity of
0.95 (95% CI, 0.94-0.96) for EUS but did not aim to
compare it to MRCP.11 Most of the included studies
were methodologically poor.

986 GASTROINTESTINAL ENDOSCOPY Volume 86, No. 6 : 2017 www.giejournal.org

Meeralam et al EUS compared with MRCP in detecting choledocholithiasis

Published reports provide a wide range of sensitivity

135 records 5160 additional
and specificity for EUS and MRCP in the diagnosis of chol- identified through records identified
edocholithiasis. A recent Cochrane review included 18 database through other
studies and tried to conduct an indirect comparison of searching sources
MRCP and EUS and concluded that both interventions pro-
vide comparable accuracy.5 The author showed an average
sensitivity and specificity of 0.95 (95% CI, 0.91-0.97) and
0.97 (95% CI, 0.94-0.99) for EUS and 0.93 (95% CI, 0.87- 239 records
0.96) and 0.96 (95% CI, 0.90-0.98) for MRCP and excluded based
concluded that these 2 modalities were comparable 271 records on title and
based solely on indirect comparison. However, none of screened abstract
the included studies represented a low risk of bias, and
the authors included only 2 head-to-head studies. There-
fore, they were not able to properly compare the accuracy
of these 2 modalities in a meta-analysis of diagnostic test 27 full-text articles
accuracy (DTA). excluded
Meta-analysis of DTA is a new method developed by
- 2 insufficient data
the Cochrane group to specifically analyze and compare
the accuracy, sensitivity, and specificity of a diagnostic - 2 not
test as compared with the reference standard.12 The head-to-head
methodology of a meta-analysis of DTA is different from comparison
a conventional meta-analysis in several aspects including -23 Inappropriate
statistical analysis and quality assessment of the included 32 full-text articles or no reference
trials. To our knowledge, a proper meta-analysis of DTA assessed standard
comparing EUS and MRCP in detecting choledocholithiasis
is lacking, and hence we aimed at determining and
comparing the diagnostic accuracy, sensitivity, and speci-
ficity of EUS and MRCP in diagnosing choledocholithiasis 5 studies included
by using recommended earlier-mentioned methodology. in synthesis

METHODS Figure 1. PRISMA (Transparent Reporting of Systematic Reviews and

Meta-Analyses) diagram of included study trials.

Registration
The study protocol was registered (CRD42017057669)
in the International prospective register of systematic re-
views (PROSPERO).
Literature search
Two individual investigators independently conducted a
comprehensive computerized medical literature search by
using OVID MEDLINE (1946 to January 2017), EMBASE
(1980 to January 2017), Cochrane library, clinical trials data-
base (www.clinicaltrials.gov), and ISI Web of knowledge
from 1980 to January 2017. Other available sources were
searched through cross-referencing. Articles were selected
by using a highly sensitive search strategy with a combina-
tion of MeSH headings and text words that included (1)
choledocholithiasis, (2) EUS, endoscopic ultrasound or en-
dosonography, and (3) MRCP or magnetic resonance imag-
ing/cholangiopancreatography. Recursive searches and
cross-referencing were carried out by using a “similar arti-
cles” function. Bibliography of the articles identified after
an initial search were manually reviewed. Data extraction
and quality control were independently done by 2 re-
viewers. A third reviewer was involved if a conflict
occurred. Corresponding authors were contacted to obtain
missing data, where appropriate, before exclusion. If
missing data could not be obtained, the trial was excluded,
and the reason was described. Studies also were excluded
if they did not use a reference standard.
Inclusion criteria and study characteristics
Only head-to-head prospective studies comparing EUS
and MRCP in detecting choledocholithiasis in patients un-
dergoing both investigations were included. Patients were
aged >18 years and had a suspicion of choledocholithiasis
based on clinical presentation, history, physical examina-
tion, and abdominal US.
Radial or linear EUS and MRCP were considered as index
tests. Accepted reference standards included ERCP or IOC
as well as clinical follow-up for at least 3 months in negative
cases. True positive, false positive, true negative, and false
negative were calculated for both EUS and MRCP.
Exclusion criteria
Studies with insufficient data, abstracts, pediatric
studies, duplicate publications, studies with no reference

www.giejournal.org Volume 86, No. 6 : 2017 GASTROINTESTINAL ENDOSCOPY 987

EUS compared with MRCP in detecting choledocholithiasis Meeralam et al

TABLE 1. Characteristics of included studies

Intervention-reference
Study Patients Reference standard standard gap EUS-MRCP gap Blinding

Fernandez-Esparrach 135 suspected CBD stones: dilated ERCP, IOC or Unclear 24 h Blinded for index
et al36 CBD or clinical presentation 6-mo FU and reference
tests

Kondo et al37 28 suspected CBD stones ERCP  IDUS <2 wk Unclear Blinded for index
and reference
tests

Aubé et al38 47 suspected CBD stones ERCP, IOC or clinical Unclear <48 h Blinded for index
and negative US FU for 3 mo tests
de Lédinghen 32 suspected CBD stones (clinical ERCP or IOC Unclear Unclear Blinded for index
et al39 and laboratory evidence) tests

Scheiman et al40 30 suspected CBD stones (clinical, ERCP 24 h 24 h Blinded to test

laboratory, and US evidence) results

All studies were prospective cohort studies.

TP, True positive; FP, false positive; FN, false negative; TN, true negative; CBD, common bile duct; IOC, intraoperative cholangiography; FU, follow-up; IDUS, intraductal US.

standards, and >48 hours’ gap between 2 index tests were
excluded. There was no restriction in terms of language,
location, or quality of the studies.
Risk of bias
The risk of bias in the included studies12,13 was evaluated
by using the QUADAS-2 tool,13 which is the
recommendation by the Cochrane Collaboration. This tool
comprises 4 domains including patient selection, index
test, reference standard, and flow of patients through the
study and timing of the index tests. It is completed in 4
phases: report the review question, develop review-
specific guidance, review the published flow diagram for
the primary study or construct a flow diagram if none is
reported, and judge bias and applicability.13
Outcome measure
The primary object of this study was to compare the
diagnostic test accuracy of EUS and MRCP as compared
with the reference standard in detecting choledocholithia-
sis by using a meta-analysis of DTA. The secondary objec-
tives of the study were to compare sensitivity and
specificity of EUS and MRCP. We also aimed to compute
the diagnostic odds ratio (DOR) and positive and negative
likelihood ratios and to investigate the source of heteroge-
neity in the final analysis, based on the methodology or
included studies.
ease.12 The DOR was defined as a single number that
described how many times higher the odds were of
obtaining a positive test result in a case with the disease
rather than a case without the disease and that
summarized the diagnostic accuracy of the index test.
Youden’s Index was computed as a general index of test
accuracy by using a combination of sensitivity and
specificity as sensitivity þ specificity – 1.12 A summary
receiver operating characteristic curve (sROC) was
developed, and an area under the curve (AUC) was
calculated. A perfect test has an AUC close to 1, and
poor tests have AUCs close to 0.5. We used a bivariate
hierarchical model to perform the meta-analysis of DTA.
To compare the accuracy of the tests, we added a covariate
for test type to the bivariate model to assess its effect on
sensitivity and specificity. The statistical significance of
the difference in test performance was assessed by using
a likelihood ratio test comparing models with and without
the covariate terms. Data analysis was performed by using
RevMan version 5.3, STATA version 12 (College Station,
Tex) and Meta-DiSc version 1.4 (Madrid, Spain). We
planned to assess the risk of publication bias by using a
funnel plot if the number of included studies was >10.12
The Moses-Shapiro-Littenberg meta-regression method
was used to explore sources of heterogeneity in the studies
by using a weighted least squares method and was pre-
sented by a relative DOR and P value.14

Statistical analysis RESULTS

The sensitivity was defined as the probability that the in-
dex test result would be positive in a case with the disease. Characteristics of included studies
The specificity was defined as the probability that the index Five of 32 studies were included. Figure 1 depicts the
test result would be negative in a case without the dis- PRISMA (Transparent Reporting of Systematic Reviews

988 GASTROINTESTINAL ENDOSCOPY Volume 86, No. 6 : 2017 www.giejournal.org

Meeralam et al EUS compared with MRCP in detecting choledocholithiasis

TABLE 1. Continued
EUS MRCP
Exclusion TP FP FN TN Sensitivity Specificity TP FP FN TN Sensitivity Specificity

Contraindications to MRCP 59 9 2 52 0.97 (0.89-1.00) 0.85 (0.74-0.93) 54 2 7 69 0.89 (078-0.95) 0.97 (0.90-1.00)
and EUS, refusal or inability
to provide informed
consent, severe pancreatitis
Contraindications to CT, 24 2 0 2 1.00 (0.86-1.00) 0.50 (0.07-0.93) 21 1 3 3 0.88 (0.68-0.97) 0.75 (0.19-0.99)
MRCP, or EUS; history of
choledocholithiasis, acute
cholangitis
Contraindications to MRCP or 15 1 1 25 0.94 (0.70-1.00) 0.96 (0.80-1.00) 14 1 2 25 0.88 (0.62-0.98) 0.96 (0.80-1.00)
EUS, positive US
Alcohol abuse, taking 10 1 0 21 1.00 (0.69-1.00) 0.95 (0.77-1.00) 10 6 0 16 1.00 (0.69-1.00) 0.73 (0.50-0.89)
hepatotoxic drug, positive
serology for hepatitis B or C
Cholangitis, hypotension, or 4 1 1 22 0.80 (0.28-0.99) 0.96 (0.78-1.00) 2 1 3 22 0.40 (0.05-0.85) 0.96 (0.78-1.00)
previously documented
claustrophobia

and Meta-Analyses) diagram for inclusion of studies and
the reasons for exclusion of other studies. Three studies
were from Europe, 1 was from Japan, and 1 was from the
United States. These studies included a total of 272 pa-
tients. Table 1 shows characteristics of included studies.
Risk of bias
Figure 2 shows the risk of bias and applicability
concerns according to the QUADAS-2 tool.13 In summary,
1 study had a low risk of bias, and the rest presented
unclear risks of bias (Fig. 1). None of the included
studies presented a high risk of bias.
Comparison of EUS and MRCP
The overall DOR of EUS was significantly higher than
that of MRCP (P Z .008). Separate analyses of specificity
and sensitivity showed that this was mainly due to signifi-
cantly higher sensitivity of EUS as compared with that of
MRCP (P Z .006). The specificity was not significantly
different between 2 modalities (P Z .42). Table 2 depicts
the details of the meta-analysis of DTA. Figure 3 shows
the forest plot for sensitivity and specificity of EUS and
MRCP in individual included studies, and Figure 4 depicts
the sROC for the meta-analysis of DTA.
Meta-regression did not show significant change in over-
all estimation of sensitivity and specificity by the estimation
of risk of bias in included studies (P Z .79, relative
DOR Z 0.89 for EUS arms and 0.99 for MRCP arms, relative
DOR Z .99).
DISCUSSION
To our knowledge, this is the first meta-analysis of DTA
comparing the diagnostic accuracy of EUS and MRCP in de-
tecting choledocholithiasis by using a head-to-head com-
parison. We showed that both EUS and MRCP provide
excellent diagnostic accuracy; however, EUS provides
significantly better sensitivity as well as overall accuracy
with comparable specificity as compared with MRCP in de-
tecting choledocholithiasis.
ERCP is well-established for the evaluation and treat-
ment of pancreatobiliary disease. However, ERCP remains
a technically challenging procedure, with failure rates
ranging from 3% to 10%, and it is highly dependent on
operator skill and experience.15-17 Hence, EUS and MRCP
have been proposed as safer alternatives before a decision
on ERCP is made. There is a scarcity of literature
comparing these 2 modalities, although much information
is available, showing that both provide reasonable diag-
nostic accuracy. For instance, a systematic review and
meta-analysis including 27 studies showed a sensitivity of
0.94 (range, 0.93-0.96) and specificity of 0.96 (range,
0.78-1.00) for EUS in patients with suspected choledocho-
lithiasis.11 Although significant heterogeneity was
observed, EUS showed excellent diagnostic accuracy,
with a positive LR of 22.41 and negative LR of 0.09. EUS
offered high sensitivity even in small stones <5 mm.11
A few systematic reviews have attempted to compare
the diagnostic accuracy of EUS and MRCP in diagnosing
choledocholithiasis. One systematic review of 5 trials
showed an aggregated sensitivity of 0.93 and 0.95 and
aggregated specificity of 0.96 and 0.95 for EUS and
MRCP, respectively.18 The authors concluded that there
was no statistical difference between the 2 groups.
However, this review suffered from several methodologic
imperfections. First, the authors claimed to include
randomized prospective trials, whereas none of the
included studies was randomized. In fact, one can argue

www.giejournal.org Volume 86, No. 6 : 2017 GASTROINTESTINAL ENDOSCOPY 989

EUS compared with MRCP in detecting choledocholithiasis Meeralam et al

Patient Selection
Index Test
Reference Standard
Flow and Timing

0% 25% 50% 75% 100% 0% 25% 50% 75% 100%

Risk of Bias Applicability Concerns
High Unclear Low

Risk of Bias Applicability Concerns

Reference Standard

Reference Standard
Patient Selection

Patient Selection
Flow and Timing
Index Test

Index Test
Aube 2003
de Ledinghen 1999
Fernandez-Esparrach 2007

Kondo 2005
Scheiman 2001

High Unclear Low

Figure 2. Risk of bias and applicability concerns summary.

TABLE 2. Summary of results of DTA meta-analysis

EUS MRCP
Heterogeneity Heterogeneity EUS vs MRCP
I2 P value I2 P value P value

DOR 162.5 (54.0-489.3) 0.0% .88 79.0 (23.8-262.2) 22.3% .27 .008
Sensitivity 0.97 (0.91-0.99) 15.1% .32 0.87 (0.80-0.93) 55.5% .06 .006
Specificity 0.90 (0.83-0.94) 54.2% .06 0.92 (0.87-0.96) 68.8% .01 .42
Positive LR 7.54 (3.02-18.83) 65.2% .02 8.99 (2.39-33.8) 76.4% .002 N/A
Negative LR 0.07 (0.03-0.15) 0.0% .65 0.19 (0.07-0.49) 74.3% .004 N/A
Youden index 0.86 NA 0.80 N/A N/A
DTA, Diagnostic test accuracy; DOR, diagnostic odds ratio; LR, likelihood ratio; N/A, not applicable.

that, given the nature of question of the study, a head-to-head comparisons. However, the authors included
randomized methodology may not be appropriate. On only 2 head-to-head trials and excluded 3 of our included
the other hand, the investigators simply calculated the studies due to insufficient data and therefore did not
average pooled sensitivity and specificity of EUS and perform a meta-analysis of those studies.5 They
MRCP, rather than using the appropriate recommended separately analyzed the studies on EUS and MRCP and
methodology to conduct a meta-analysis of DTA. concluded that both EUS and MRCP provide reasonable
The authors also combined all other benign etiologies, diagnostic accuracy based on indirect comparison.
including chronic pancreatitis, that may introduce bias. Another systematic review included 8 studies, but none
To our knowledge, a Cochrane review is probably the of them performed a direct comparison of EUS and
closest attempt to perform a meta-analysis of DTA of the MRCP. The authors presented an average sensitivity and

990 GASTROINTESTINAL ENDOSCOPY Volume 86, No. 6 : 2017 www.giejournal.org

Meeralam et al EUS compared with MRCP in detecting choledocholithiasis

0.001 1 1000.1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Negative LR Specificity Sensitivity Specificity

0.001 1 1000. 0.001 1 1000.0 0.001 1 1000. 0.001 1 1000.

Positive LR Negative LR Positive LR Negative LR
Endoscopic Ultrasound Magnetic Resonance Cholangiopancreatography

Figure 3. The forest plot for sensitivity and specificity of EUS and MRCP.

specificity of the 2 tests across studies and concluded that

EUS provides higher numerical sensitivity, positive
predictive value, and negative predictive value without
presenting a statistical comparison.19 The authors visually
compared the 2 ROC curves and presented a simple
average sensitivity and specificity of 2 tests across
studies. They included 3 additional references as
compared with our meta-analysis as follows: One included
study provided overall accuracy of EUS and MRCP for
several indications in addition to choledocholithiasis but
did not provide raw data for false positive or true negative
results in patients with choledocholithiasis, and this data
could not be calculated based on presented results.20
This study was therefore excluded.
Another study used bile duct dilation or obstruction
as outcome rather than choledocholithiasis and did not
provide raw data on diagnostic accuracy of the 2 modalities
in detecting choledocholithiasis and therefore could not
be used.21 A third study was excluded because of a
methodologic reason, given that they allowed up to a
5-day interval between EUS and MRCP, and they followed
negative results as short a time as 17 days, rather than the
Figure 4. The summary receiver operating characteristic curves for sensi-
3 months required in our review.22 Another systematic tivity and specificity of EUS and MRCP. LR, likelihood ratio.
review used a conventional meta-analytic method to
compare the sensitivity and specificity of EUS and MRCP.23
The author showed similar sensitivity and specificity for stone size.24 Unfortunately, we did not have sufficient
EUS and MRCP based on this analysis. However, this data to perform subgroup analysis based on the size of
methodology is not appropriate nor recommended, and it the stone.
is not designed to compare diagnostic accuracy. Our finding might have clinical implications in the cur-
Both EUS and MRCP are operator-dependent, and the rent practice of diagnosis and management of choledocho-
interpretation of findings is subject to bias. However, the lithiasis. The current guideline by the American Society for
observed superiority in the sensitivity of EUS as compared Gastrointestinal Endoscopy recommends both EUS and
with MRCP might be due to better accuracy of EUS in MRCP as non-surgical modalities for diagnosing choledo-
detection of small stones. Some evidence suggests that cholithiasis in symptomatic patients with cholelithiasis
the accuracy of EUS is independent of stone size, whereas and an intermediate risk of disease or in patients who
the sensitivities of an MRCP or an ERCP decrease with have undergone cholecystectomy when initial laboratory

www.giejournal.org Volume 86, No. 6 : 2017 GASTROINTESTINAL ENDOSCOPY 991

EUS compared with MRCP in detecting choledocholithiasis
and US data are abnormal but nondiagnostic.25 However,
the accuracy of these guidelines was low in a cohort of
498 patients (62.1%; sensitivity, 46.7%; and specificity,
73%), which led to overutilization of ERCP.26 Based on
our results and, if appropriate, in the clinical context,
EUS may be the test of choice in these patients, given
higher diagnostic accuracy and the fact that its sensitivity
is unlikely to be affected by the size of the stone.27 On
the other hand, EUS-directed ERCPs have been advocated
as a cost-effective method in patients with high or interme-
diate risk for choledocholithiasis in several studies.28-31
These studies showed an excellent negative predictive
value for EUS before ERCP, likely because of higher sensi-
tivity consistent with our findings. The results of our study
could hence emphasize more targeted ERCP procedures,
given higher sensitivity and the fact that both EUS and
ERCP could be done in the same session.28-30
The results of our meta-analysis should be interpreted
with caution, given the limitations in conducting a meta-
analysis. Our study included only 5 studies. A meta-
analysis involving more studies and a higher number of
included patients would likely provide a more accurate es-
timate or comparison of results. We cannot completely
exclude the possibility of publication bias, but we tried to
minimize the chance by avoiding restriction in the lan-
guage of publications or quality of the included studies
and by using several search strategies. On the other
hand, we have encountered heterogeneity in choosing
the reference standard in each of the included studies.
The reference standard ranged from ERCP and IOC to clin-
ical follow-up for negative results. Although this is more
consistent with real-life clinical practice, it could cause veri-
fication bias, and we had no means, based on the available
data, to exclude this.
Furthermore, the assumption of 100% accuracy in a
reference standard does not really hold true in clinical
practice, including the case of choledocholithiasis. Small
stones could be missed by both ERCP and IOC.32 Small
air bubbles can be mistaken for stones. This might
underestimate the diagnostic accuracy of both EUS and
MRCP. The sample size of most included studies was
under 50. The accuracy of the results could have been
improved by including a larger number of patients.
However, in the absence of large studies, meta-analyses
are an optimal method of synthesizing the data from avail-
able studies. The observed limited number of studies on
this topic might be based on several factors. First, it is diffi-
cult and it may not be ethical to recruit patients to undergo
3 different invasive or semi-invasive tests for a relatively
benign condition. In addition, institutional review boards
might lose interest in approving a protocol if a few studies
have been done on the topic. However, the results of this
study are still robust, given the strict inclusion criteria to
ensure reasonable methodologic quality of the included
studies as well as the appropriate methodology to compare
the 2 modalities.
992 GASTROINTESTINAL ENDOSCOPY Volume 86, No. 6 : 2017
Meeralam et al
Moreover, the decision regarding the best modality in
an individual patient usually involves the cost of the pro-
cedure. MRCP has been slightly more costly than EUS.33
In a decision analysis on the role of EUS, ERCP, and IOC
in patients suspected for choledocholithiasis,
investigators showed that although patients with high-
risk pretest probabilities of choledocholithiasis (>50%)
may benefit from ERCP, those with a moderate pretest
probability (10%-50%) will be best served by EUS. Further-
more, a cost-effectiveness analysis demonstrated that lapa-
roscopic cholecystectomy with routine IOC is likely the
preferred strategy in the management of symptomatic
cholelithiasis with asymptomatic choledocholithiasis. How-
ever, the authors did not investigate the role of EUS or
MRCP in patients with choledocholithiasis without symp-
tomatic cholelithiasis.34 A retrospective review of medical
records in 527 patients with choledocholithiasis
compared MRCP with ERCP and IOC. Among patients
undergoing MRCP as the initial procedure, 82%
subsequently underwent either ERCP or laparoscopic
cholecystectomy. In patients undergoing initial MRCP
followed by ERCP or IOC, the sensitivity and specificity
of the test was 0.90 and 0.86. Most of the time, MRCP
was followed with a more-invasive test. The authors
concluded that MRCP did not change the management of
patients with suspected choledocholithiasis, and they
questioned its utility in this patient population.35
In conclusion, both EUS and MRCP provide excellent
diagnostic accuracy, with EUS providing statistically signif-
icantly higher diagnostic accuracy and sensitivity, most
likely due to a higher detection rate of small choledocholi-
thiasis. EUS should be incorporated into the diagnostic al-
gorithm in patients suspected for choledocholithiasis
whenever appropriate, given its reasonable safety profile.
This might apply to the patients who need an EGD for
investigating other alternative causes of abdominal pain.
Future head-to-head studies should focus on the size or
clinical significance of the choledocholithiasis identified
by each modality.
REFERENCES
1. Anciaux ML, Pelletier G, Attali P, et al. Prospective study of clinical and
biochemical features of symptomatic choledocholithiasis. Dig Dis Sci
1986;31:449-53.
2. Roston AD, Jacobson IM. Evaluation of the pattern of liver tests and
yield of cholangiography in symptomatic choledocholithiasis: a pro-
spective study. Gastrointest Endosc 1997;45:394-9.
3. Barbara L, Sama C, Morselli Labate AM, et al. A population based study
on the prevalence of gallstone disease: the Sirmione study. Hepatol-
ogy 1987;7:913-7.
4. Loria P, Dilengite MA, Bozzzoli M, et al. Prevalence rates of gallstone disease
in Italy. The Chianciano population study. Eur J Epidemiol 1994;10:143-50.
5. Giljaca V, Gurusamy KS, Takwoingi Y, et al. Endoscopic ultrasound
versus magnetic resonance cholangiopancreatography for com-
mon bile duct stones. Cochrane Database Syst Rev 2015;2:
CD011549.
www.giejournal.org
Meeralam et al
6. Barkun AN, Barkun JS, Fried GM, et al. Useful predictors of bile duct
stones in patients undergoing laparoscopic cholecystectomy. Ann
Surg 1994;220:32-9.
7. Ripolles T, Ramirez-Fuentes C, Martinez-Perez MJ, et al. Tissue har-
monic sonography in the diagnosis of common bile duct stones: a
comparison with endoscopic retrograde cholangiography. J Clin Ultra-
sound 2009;37:501-6.
8. Stiris MG, Tennoe B, Aadland E, et al. MR cholangiopancreaticography and
endoscopic retrograde cholangiopancreaticography in patients with sus-
pected common bile duct stones. Acta Radiologica 2000;41:269-72.
9. Freeman ML, Nelson DB, Sherman S, et al. Complications of endo-
scopic biliary sphincterotomy. N Engl J Med 1996;335:909-18.
10. Schofl R, Haefner M. Diagnostic cholangiopancreatography. Endoscopy
2003;35:145-55.
11. Tse F, Liu L, Barkun AN, et al. EUS: a meta-analysis of test performance
in suspected choledocholithiasis. Gastrointest Endosc 2008;67:235-44.
12. Macaskill P, Gatsonis C, Deeks JJ, et al. Analysing and presenting re-
sults. In: Deeks JJ, Bossuyt PM, Gatsonis C, editors. Cochrane handbook
for systematic reviews of diagnostic test accuracy, version 1.0. The
Cochrane Collaboration, 2010. Available at: http://srdta.cochrane.org/.
Accessed September 12, 2016.
13. Whiting PF, Rutjes AW, Westwood ME, et al; QUADAS-2 Group. QUA-
DAS-2: a revised tool for the quality assessment of diagnostic accuracy
studies. Ann Intern Med 2011;155:529-36.
14. Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heteroge-
neity in systematic reviews of diagnostic tests. Stat Med 2002;21:
1525-37.
15. Fulcher AS. MRCP and ERCP in the diagnosis of common bile duct
stones. Gastrointest Endosc 2002;56(suppl 6):S178-82.
16. Bilbao MK, Dotter CT, Lee TG, et al. Complications of endoscopic retro-
grade cholangiopancreatography (ERCP): a study of 10,000 cases.
Gastroenterology 1976;70:314-20.
17. Cotton PB, Yaghoobi M. Adverse events. In: Cotton PB, Leung J, editors.
ERCP: The fundamentals, 2nd ed. Chichester (UK): John Wiley & sons;
2014. p. 369-96.
18. Verma D, Kapadia A, Eisen GM, et al. EUS vs MRCP for detection of
choledocholithiasis. Gastrointest Endosc 2006;64:248-54.
19. De Castro VL, Moura EG, Chaves DM, et al. Endoscopic ultrasound versus
magnetic resonance cholangiopancreatography in suspected choledo-
cholithiasis: a systematic review. Endosc Ultrasound 2016;5:118-28.
20. Ainsworth AP, Rafaelsen SR, Wamberg PA, et al. Is there a difference in
diagnostic accuracy and clinical impact between endoscopic ultraso-
nography and magnetic resonance cholangiopancreatography?
Endoscopy 2003;35:1029-32.
21. Materne R, Van Beers BE, Gigot JF, et al. Extrahepatic biliary obstruc-
tion: magnetic resonance imaging compared with endoscopic ultraso-
nography. Endoscopy 2000;32:3-9.
22. Schmidt S, Chevallier P, Novellas S, et al. Choledocholithiasis: repetitive
thick-slab single-shot projection magnetic resonance cholangiopan-
creaticography versus endoscopic ultrasonography. Eur Radiol
2007;17:241-50.
23. Ledro-Cano D. Suspected choledocholithiasis: endoscopic ultrasound
or magnetic resonance cholangio-pancreatography? A systematic re-
view. Eur J Gastroenterol Hepatol 2007;19:1007-11.
24. Aljebreen A, Azzam N, Eloubeidi MA. Prospective study of endoscopic
ultrasound performance in suspected choledocholithiasis.
J Gastroenterol Hepatol 2008;23:741-5.
25. ASGE Standards of Practice Committee; Maple JT, Ben-Menachem T,
Anderson MA, et al. The role of endoscopy in the evaluation of sus-
pected choledocholithiasis. Gastrointest Endosc 2010;71:1-9.
26. Adams MA, Hosmer AE, Wamsteker EJ, et al. Predicting the likelihood
of a persistent bile duct stone in patients with suspected choledocho-
www.giejournal.org
EUS compared with MRCP in detecting choledocholithiasis
lithiasis: accuracy of existing guidelines and the impact of laboratory
trends. Gastrointest Endosc 2015;82:88-93.
27. Sugiyama M, Atomi Y. Endoscopic ultrasonography for diagnosing
choledocholithiasis: a prospective comparative study with ultrasonog-
raphy and computed tomography. Gastrointest Endosc 1997;45:143-6.
28. Lee YT, Chan FKL, Leung WK, et al. Comparison of EUS and ERCP in the
investigation with suspected biliary obstruction caused by choledo-
cholithiasis: a randomized study. Gastrointest Endosc 2008;67:660-8.
29. Liu CL, Fan ST, Lo CM, et al. Comparison of early endoscopic ultraso-
nography and endoscopic retrograde cholangiopancreatography in
the management of acute biliary pancreatitis: a prospective random-
ized study. Clin Gastroenterol Hepatol 2005;3:1238-44.
30. Karakan T, Cindoruk M, Alagozlu H, et al. EUS versus endoscopic retro-
grade cholangiography for patients with intermediate probability of
bile duct stones: a prospective randomized trial. Gastrointest Endosc
2009;69:244-52.
31. Buscarini E, Tansini P, Vallisa D, et al. EUS for suspected choledocholi-
thiasis: Do benefits outweigh costs? A prospective controlled study.
Gastrointest Endosc 2003;57:510-8.
32. Liu CL, Lo CM, Chan JK, et al. EUS for detection of occult cholelithiasis
in patients with idiopathic pancreatitis. Gastrointest Endosc 2000;51:
28-32.
33. Sahai AV, Mauldin PD, Marsi V, et al. Bile duct stones and laparoscopic
cholecystectomy: a decision analysis to assess the roles of intraopera-
tive cholangiography, EUS, and ERCP. Gastrointest Endosc 1999;49:
334-43.
34. Epelboym I, Winner M, Allendorf JD. MRCP is not a cost-effective strat-
egy in the management of silent common bile duct stones.
J Gastrointest Surg 2013;17:863-71.
35. Badger WR, Borgert AJ, Kallies KJ, et al. Utility of MRCP in clinical deci-
sion making of suspected choledocholithiasis: an institutional analysis
and literature review. Am J Surg. Epub 2016 Dec 1.
36. Fernández-Esparrach G, Ginès A, Sánchez M, et al. Comparison of
endoscopic ultrasonography and magnetic resonance cholangiopan-
creatography in the diagnosis of pancreatobiliary diseases: a prospec-
tive study. Am J Gastroenterol 2007;102:1632-9.
37. Kondo S, Isayama H, Akahane M, et al. Detection of common bile duct
stones: comparison between endoscopic ultrasonography, magnetic
resonance cholangiography, and helical-computed-tomographic chol-
angiography. Eur J Radiol 2005;54:271-5.
38. Aubé C, Delorme B, Yzet T, et al. MR cholangiopancreatography versus
endoscopic sonography in suspected common bile duct lithiasis: a
prospective, comparative study. AJR Am J Roentgenol 2005;184:55-62.
39. de Lédinghen V, Lecesne R, Raymond JM, et al. Diagnosis of choledo-
cholithiasis: EUS or magnetic resonance cholangiography? A prospec-
tive controlled study. Gastrointest Endosc 1999;49:26-31.
40. Scheiman JM, Carlos RC, Barnett JL, et al. Can endoscopic ultrasound or
magnetic resonance cholangiopancreatography replace ERCP in pa-
tients with suspected biliary disease? A prospective trial and cost anal-
ysis. Am J Gastroenterol 2001;96:2900-4.
Received March 6, 2017. Accepted June 12, 2017.
Current affiliation: Division of Gastroenterology, McMaster University,
Hamilton, Ontario, Canada.
Reprint requests: Mohammad Yaghoobi, MD, MSc (Epi), AFS, DABIM,
FRCPC, Division of Gastroenterology, McMaster University, 1280 Main
Street West, Rm-2F, Hamilton ON, L8S 4K1 Canada.
Volume 86, No. 6 : 2017 GASTROINTESTINAL ENDOSCOPY 993