3 Results and recommendations
A summary of all recommendations is provided in [Table 2].
Table 2 ESGE recommendations on the diagnostic work-up of biliary strictures.
|
Recommendations
|
Strength of recommendation
|
Quality of evidence
|
|
CECT, contrast-enhanced computed tomography; ERCP, endoscopic retrograde
cholangiopancreatography; ERCP-TA, ERCP-based tissue acquisition; EUS, endoscopic
ultrasound; EUS-TA, EUS-guided tissue acquisition; MRCP, magnetic resonance
cholangiopancreatography; MRI, magnetic resonance imaging.
|
|
1 ESGE suggests against the use of tumor markers alone to
discriminate between malignant and benign causes of a stricture in patients with
suspected biliary stricture presenting with jaundice; ESGE suggests the use of
laboratory tests in addition to other investigations (imaging and histopathology)
in
making a definite diagnosis
|
Conditional
|
Very low
|
|
2 ESGE suggests the use of MRI/MRCP over CECT to discriminate between malignant and
benign cause of obstruction and to detect the level of the stricture in patients with
suspected biliary stricture presenting with jaundice and/or biochemical evidence of
cholestasis
|
Conditional
|
Low
|
|
3 ESGE recommends the combination of EUS-TA and ERCP-TA
as the preferred diagnostic approach for tissue acquisition in patients with jaundice
and distal extrahepatic biliary stricture in the absence of a pancreatic mass
|
Strong
|
Moderate
|
|
4 ESGE suggests performing EUS and ERCP in the same session as this approach does not
increase the risk of adverse events nor impair the effectiveness of the techniques
|
Conditional
|
Very Low
|
|
5 ESGE suggests performing EUS in patients with suspected biliary stricture and/or
unexplained common bile duct dilatation even in the absence of abnormal laboratory
tests
|
Conditional
|
Very Low
|
|
6 ESGE suggests that brushing cytology should be completed along with fluoroscopy-guided
biopsies, wherever technically feasible, in patients with perihilar biliary strictures
|
Conditional
|
Low
|
|
7 ESGE suggests EUS-TA for perihilar strictures when ERCP-based modalities yield insufficient
results, provided that curative resection is not feasible and/or when cross-sectional
imaging has shown accessible extraluminal disease
|
Conditional
|
Low
|
|
8 ESGE suggests using standard ERCP diagnostic modalities
at index ERCP. In the case of indeterminate biliary strictures, ESGE suggests taking
cholangioscopy-guided biopsies, in addition to standard ERCP diagnostic modalities.
Additional intraductal biliary imaging modalities can be selectively used, based on
clinical context, local expertise, and resource availability
|
Conditional
|
Very Low
|
The presence of a biliary stricture portends a medical challenge for any endoscopist,
in
terms of its anatomy and difficulty in making a diagnosis, and subsequently in deciding
on its
management [4]. Differentiating whether a biliary stricture is either benign or malignant is of
the
foremost importance as it has implications in surgical and oncologic decision-making,
hence
the need to make a safe, accurate, and expedient diagnosis [5]. The gold standard has always been tissue diagnosis, wherein lies the role of
endoscopy (ERCP or EUS) in providing cytologic or histologic specimens. Among the
most common
benign causes of biliary stricture is iatrogenic bile duct injury, whereas among malignant
causes, pancreatic ductal adenocarcinoma (PDAC) and cholangiocarcinoma (CCA) are the
commonest
causes of distal and proximal stricture, respectively [6]. A list of the main etiologies of biliary stricture is shown in the list next page.
Sometimes biliary strictures can present without an underlying mass, which occasionally
makes
them unidentifiable on cross-sectional imaging, as well as limiting the utility of
tissue
diagnosis in view of the desmoplastic nature of the tumors [7]
[8]
[9].
Main etiologies of biliary strictures
Malignant
-
Pancreatic cancer
-
Cholangiocarcinoma
-
Gallbladder cancer
-
Hepatocellular carcinoma
-
Ampullary cancer
-
Lymphoma
-
Rare neoplasia (cystadenocarcinoma, mixed hepatocellular-cholangiocellular
cancer)
-
Metastasis (colon cancer, breast cancer, renal cell cancer)
Fibroinflammatory
-
Chronic pancreatitis
-
Primary sclerosing cholangitis
-
Autoimmune pancreatitis
-
IgG4-mediated cholangitis
-
Sarcoidosis
-
Recurrent pyogenic cholangitis
-
Extrinsic compression by pancreatic fluid collections
Iatrogenic
-
Cholecystectomy
-
Liver transplantation
-
Local cancer treatment
Vascular
Other
-
AIDS cholangiopathy
-
Mirizzi syndrome
AIDS, acquired immunodeficiency syndrome; IgG4, immunoglobulin G4.
3.1 Laboratory tests and cross-sectional imaging
ESGE suggests against the use of tumor markers alone to discriminate between malignant
and benign causes of a stricture in patients with suspected biliary stricture presenting
with jaundice; ESGE suggests the use of laboratory tests in addition to other investigations
(imaging and histopathology) in making a definite diagnosis.
Conditional recommendation, very low quality of evidence.
Inexpensive and readily available serologic tests that can rapidly differentiate between
benign and malignant causes of biliary stricture would allow prioritization of patients.
The two most commonly studied tumor markers related to the pancreaticobiliary system
are the carbohydrate antigen 19–9 (CA19–9) and carcinoembryonic antigen (CEA). Both
are glycoprotein tumor markers with the immune-determinant epitope on the carbohydrate
and protein moiety of the molecule, respectively [10]
[11].
3.1.1 Benefits
The usefulness of these tumor markers has been shown in multiple studies, regarding
their role in staging, evaluation of resectability status, and assessment of prognosis
and of recurrence on follow-up [12]
[13]
[14]
[15]
[16].
The panel conducted a comprehensive meta-analysis to determine the accuracy of CA19–9
and CEA levels in differentiating benign versus malignant etiologies. After a comprehensive
literature search, 11 studies for CA19–9 [17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27] and four studies for CEA levels [26]
[28]
[29]
[30] were selected for analysis (inclusion criteria: CA19–9 >37 U/mL and CEA >2.5–5 ng/mL,
with studies reporting the accuracy of malignant vs. benign cases) (Table 3s). The pooled accuracy of CA19–9 levels for discrimination between benign and malignant
causes of biliary stricture was 81% (95%CI 76%–87%; I
2=84.7%) (Fig. 1s, part a), whereas the pooled accuracy of CEA levels was 70% (95%CI 62%–78%; I
2=63.7%) (Fig. 1s, part b). For malignant cases, the sensitivity of CA19–9 levels was 68%–85%, with a specificity
of 64%–87%. Similarly, for CEA levels, the sensitivity and specificity were 25%–82%
and 83%–92%, respectively.
Morris-Stiff et al. have also reported that the combination of tumor markers (CA19–9)
with cross-sectional imaging increases the sensitivity and specificity from 85% to
97% and 70% to 89%, respectively [21]. Similarly, the decline in CA19–9 post-biliary drainage is slower, and the average
reduction is smaller in malignant cases (50%) compared with benign cases (100%) [17]. All these factors highlight that the role of tumor markers (especially CA19–9 and
CEA) is complementary to other investigations, like cross-sectional imaging and tissue
diagnosis, for determining the underlying etiology of biliary strictures.
Additionally, when IgG4-related sclerosing cholangitis (IgG4-SC) is clinically suspected,
especially in a setting of distal common bile duct (CBD) stricture and autoimmune
pancreatitis, serologic testing in the form of serum IgG4 levels is usually the first-line
investigation, having been reported to be elevated in 90% of IgG4-SC cases [31]
[32].
3.1.2 Harms
Numerous studies have reported elevated CA19–9 levels (>37 U/mL) in various
pancreaticobiliary cancers (especially PDAC, CCA, and gallbladder cancer). CA19–9
proved a
more useful marker in the absence of cholestasis or cholangitis (sensitivity 78% vs.
74%;
specificity 83% vs. 42%) [33]. Increasing the cutoff value decreases the sensitivity and precludes its clinical
utility. CA19–9 has also been shown to be elevated in a number of benign conditions,
such
as liver diseases (primary sclerosing cholangitis [PSC] and cirrhosis, among others),
obstructive jaundice, pancreatitis, and even in renal, respiratory, and rheumatologic
cases [34]. Moreover, CA19–9 antigen is an oligosaccharide corresponding to sialylated Lewis
blood group Lea
, whose concentration in serum may be
influenced by the patient’s secretor status and Lewis genotype [35] (7% of the world’s population are nonsecretors). As a result, such false-positive
cases are to be interpreted with caution.
As evident in the meta-analysis above, both tumor markers have low pooled accuracy
in identifying malignancy as a cause of biliary obstruction. Moreover, high false-positive
rates of CA19–9 and CEA have been reported (15% and 18%, respectively) [19]. Higher cutoff values for CA19–9 were suggested in a few studies (e.g. 90 U/mL and
70.5 U/mL), which further decreased the sensitivity [17]
[21].
Although the serum IgG4 level is a piece of the diagnostic evidence when IgG4-SC is
suspected, its specificity is suboptimal to be used in isolation, without other criteria.
Kamisawa et al. reported using a cutoff of >135 mg/dL along with both imaging and
histopathologic evidence to formulate a definitive diagnosis [36]. Its sensitivity and specificity are 64%–90% and 87%–93%, respectively [31]
[36]
[37].
Elevated levels of serum IgG4 have also been reported in patients with CCA (8%–14%
of
patients), PSC (9%–22%), and pancreatic cancer (6%) [31]
[37]
[38]
[39]. Therefore, owing to the low discrimination value of serum IgG4 levels when used
alone and frequent false positives in various diseases, especially PSC and CCA (which
may
mimic IgG4-SC), IgG4 levels should not be used alone for the diagnosis of IgG4-SC,
but
only in combination with imaging and histopathology.
In conclusion, laboratory tests and serum tumor markers alone have an unacceptably
low accuracy in differentiating between benign and malignant biliary obstruction,
especially in the presence of jaundice or cholangitis.
3.1.3 Cost-effectiveness profile
Although measurement of tumor markers is relatively inexpensive, the costs associated
with searching for malignancy on the basis of an isolated finding of an elevated tumor
marker can amount to thousands of dollars, for example US$34685 for ERCP and $37458
for EUS according to a recent cost-effectiveness analysis [40]. Therefore, although specific cost-effectiveness analyses on the use of tumor markers
in biliary strictures are lacking, the unacceptably low accuracy of tumor markers
alone in detecting neoplasia would determine an increase in the costs of missed diagnoses
and an overuse of cross-sectional imaging and endoscopy in false-positive cases.
3.1.4 Environmental impact
Laboratory buildings, processes, and equipment, including single-use plastics,
automation, and technologies using vast amounts of energy and water, along with
transportation and storage of samples, all add to the environmental burden. Among
the
suggestions to reduce the carbon footprint in this setting are recycling bins in all
laboratory and nonlaboratory areas, reviewing freezer contents regularly and discarding
items not needed, performing audits to reduce printing, storing work logs digitally
to
reduce printing, nonlaboratory staff working from home on a rota basis, using timers
or
labelling to switch off nonessential equipment when not in use, and using electronic
signatures. Furthermore, reducing sample transport and finding more efficient means
of
transport represent another challenge to improve this aspect [41].
3.1.5 Certainty of the evidence assessment
The risk of bias assessment for each study can be found in Table 4s. Overall, all the included studies were considered to be of good quality. The certainty
of evidence in this PICO question was downgraded owing to the fact that there were
only non-comparative observational studies included (Table 5s). For the main outcome of diagnostic accuracy, further downgrading was applied owing
to inconsistency (high heterogeneity), which downgraded the quality of evidence to
very low.
ESGE suggests the use of MRI/MRCP over contrast-enhanced computed tomography (CECT)
to discriminate between malignant and benign causes of obstruction and to detect the
level of the stricture in patients with suspected biliary stricture presenting with
jaundice and/or biochemical evidence of cholestasis.
Conditional recommendation, low quality of evidence.
The expanding spectrum of therapeutic interventions for the management of biliary
stricture mandates that a precise assessment be made beforehand regarding the etiology,
location, level, and extent of the disease. This diagnostic algorithm usually follows
a two-pronged approach: initial clinical and laboratory examination, followed by cross-sectional
imaging. Any imaging procedure that is undertaken should answer two important questions:
etiology (benign vs. malignant) and the level of the stricture. The role of imaging
in the evaluation of any case of biliary stricture entails confirming the diagnosis
(benign vs. malignant), locating the level of stricture (distal, mid, or hilar), staging
the disease, assessing the resectability status, and giving preoperative information
to the surgeon by providing a proper roadmap and the relevant anatomic and vascular
anatomy.
Transabdominal ultrasound has low accuracy in determining the etiology and extent
of the disease, with a sensitivity and specificity for biliary obstruction of 31%–100%
and 71%–97%, respectively [42]
[43]
[44]. In contrast, ERCP (the gold standard) remains the cornerstone in the management
of biliary strictures, offering simultaneous tissue diagnosis, but is burdened with
complications (4%–10%), such as post-ERCP pancreatitis, bleeding, infection, and perforation,
meaning its role for diagnostic purposes has become outdated [42]
[43]
[45].
3.1.6 Benefits
MRCP is noninvasive, and does not require contrast injection and ionizing radiation
exposure, with minimal risk of post-procedure complications. It provides a panoramic
view of the biliary tree (multiplanar capability), both above and below the stricture,
which enables diagnosis of the site, nature, and dimensions of the obstruction, providing
a complete overview of the biliary tree to the endoscopist/surgeon. When combined
with MRI, it allows indirect visualization of drainage catheters and preoperative
evaluation of the extent of disease (staging), which is especially useful in those
with altered anatomy. Meanwhile, a contrast-enhanced computed tomography (CECT) scan
is quick, with few motion artifacts, and with good patient compliance; however, it
involves exposure to radiation and intravenous contrast agents. All these factors
give MRI/MRCP a “theoretical edge” over CECT scanning for the evaluation of any cases
of biliary stricture [6]
[46].
For the evaluation of obstructive jaundice, the published literature reports a sensitivity,
specificity, and accuracy of MRCP of 81%–100%, 84%–100%, and 90%–96%, respectively
[47]. Likewise, a CECT abdomen is highly sensitive (74%–96%) and specific (90%–94%) for
the detection of biliary obstruction [42]
[43]
[44]. Both modalities can also detect vascular encasement and metastatic disease.
The panel conducted a comprehensive meta-analysis to assess the pooled accuracy of
CECT versus MRI/MRCP in determining the level of biliary stricture and malignancy
as the etiology. Only head-to-head comparative studies were taken into consideration.
For determination of the level of stricture, six studies were analyzed [48]
[49]
[50]
[51]
[52]
[53] (Table 6s), wherein the pooled accuracy of MRI/MRCP was superior to that of CECT (odds ratio
[OR] 3.3, 95%CI 1.2–9.1; I
2=0%) (Fig. 2s, part a).
Similarly, 10 studies [47]
[49]
[50]
[51]
[53]
[54]
[55]
[56]
[57]
[58] were analyzed for the pooled accuracy of diagnosing malignant biliary strictures
(Table 7s), wherein MRI/MRCP again proved superior to CECT (96%
vs. 89%; OR 2.1, 95%CI 1.2–3.6; I
2=12%) (Fig. 2s, part b). Furthermore, MRI/MRCP has also been shown to avoid
the need for unnecessary ERCP in approximately one-third of cases [59], is valuable in the setting of failed ERCP, and offers incremental benefit over
ultrasound in pregnant patients with suspected biliary obstruction. MRI/MRCP is also
useful in planning for therapeutic ERCP in difficult situations, such as complex hilar
strictures, or where additional pancreatic therapy is being contemplated [60]
[61].
In conclusion, based on the evidence above, ESGE suggests that, in a suspected case
of biliary stricture with/without jaundice, MRI/MRCP should be preferred over CECT
for the evaluation of both the level and the underlying pathology of the biliary stricture.
3.1.7 Harms
Even though MRCP is superior to CECT in evaluating the cause and level of a biliary
stricture, both have their own sets of limitations. Both are operator-dependent modalities,
and their diagnostic accuracy varies with the available expertise [62].
Contraindications for the use of MRI/MRCP include the presence of underlying
pacemakers, cerebral aneurysm clips (although MRI-compatible devices are now available),
and claustrophobia. Furthermore, technical considerations, such as breath-holding,
lack of
therapeutic ability, and motion artifacts, limit its usefulness in certain situations.
The
latter leads to poor sensitivity of MRCP for the detection of small stones (<3–5 mm;
64%), as reported by Guibaud et al. [63]. Motion artifacts are prevalent in the distal portion of the CBD, where small
stones usually migrate, thereby decreasing resolution, which has been shown to translate
into lower performance for MRCP when evaluating distal over proximal strictures [64]. By extrapolation of these findings, ampullary pathologies may be erroneously
diagnosed because of poor performance of MRCP at/near the duodenal wall. A meta-analysis
of 67 studies by Romagnuolo et al. reported that a negative MRCP did not exclude a
diagnosis, if the pretest suspicion of malignancy was high, with suspicion only reduced
by
10% by a normal MRCP [65].
Similarly, in patients with known hypersensitivity or allergy to contrast agents,
or chronic kidney disease, CECT scanning is seldom used. The use of contrast agents
is a major limitation to the use of CECT, as CECT formulates a diagnosis of luminal
or extraluminal pathology based on enhancement patterns and attenuation value in the
portal venous phase. In addition, radiation exposure is another factor that gives
MRI/MRCP an edge over CECT for the evaluation of biliary diseases. The increasing
use of CECT has heightened concerns about relatively large cumulative exposure to
radiation owing to repeated examinations. The average typical effective doses for
a basic CT of the abdomen and the chest are 8 mSv and 7 mSv, respectively [66], although the cancer risk is negligible when examined from a population level. Extrapolation
of data from lung cancer screening clearly shows that the benefits of using CECT scanning
at present outweigh the risks associated with cumulative exposure to radiation [67], provided it is not used for asymptomatic cases.
3.1.8 Cost-effectiveness profile
Performing MRI/MRCP prior to ERCP in planning for a subsequent therapeutic strategy
has been shown to be a more cost-effective approach for biliary diseases. Howard et
al. reported, in a probabilistic sensitivity analysis, that with a low–moderate probability
of bile duct stones in a post-cholecystectomy situation, there was a 59% likelihood
of MRCP being a cost-saving procedure, with an 83% chance that MRCP had a cost-effectiveness
ratio more favorable than $50000 per quality-adjusted life year (QALY) gained [68]. Similar findings were echoed by Vergel et al., wherein MRCP was both cost-effective
and showed improved quality of life outcomes, compared with diagnostic ERCP [69].
3.1.9 Environmental impact
A recent systematic review highlighted that the energy consumption of diagnostic radiology
devices, such as MRI and CECT scanners, constitutes a significant portion of the overall
energy usage in a radiology department [70]. Interestingly, this systematic review reported that a considerable percentage (40%–91%)
of the energy consumption of these devices was defined as nonproductive, emphasizing
the need to optimize device usage and reduce idle times [70]. Other sustainability actions were shown to be implementing informatic tools for
workflow and optimizing operating modes, with estimated annual savings ranging from
14180 to 171000 kWh [70]. For MRI, the carbon footprint associated with the entire individual service has
been measured up to a maximum of 22.4 kg CO2-equivalents, as Scope 2 and Scope 3 emissions. This value includes both in-hospital
process energy (29 kWh per patient) and out-of-hospital energy (about 75 kWh per patient),
which is required not only for electricity generation during its medical use but also
for the manufacturing of the MRI scanner itself and the disposable and reusable products
used during diagnostic activities [71]. Only approximately 28% of the total MRI life-cycle energy is used for image acquisition,
suggesting potential improvements to reduce the environmental impact [71].
3.1.10 Certainty of the evidence
The risk of bias assessment for each study can be found in Table
8s. Overall, the included studies were considered to be of good quality. The
certainty of evidence for all clinical outcomes in this PICO question was downgraded
because the evidence was based only on nonrandomized observational studies in the
absence
of RCTs (Table 5s). Therefore, for both the outcomes of
diagnostic accuracy for the detection of the obstruction level and of determining
the
malignant etiology of the stricture, the quality of evidence was rated as low and
a
conditional recommendation was proposed.
3.2 Distal biliary strictures
ESGE recommends the combination of EUS-guided tissue acquisition and ERCP-based sampling
as the preferred diagnostic approach for tissue acquisition in patients with jaundice
and distal extrahepatic biliary stricture in the absence of a pancreatic mass.
Strong recommendation, moderate quality evidence.
Tissue sampling to establish the nature of the stenotic lesion is a key step in the
diagnostic work-up of biliary strictures. Although these strictures can be due to
both extrinsic compression and intrinsic pathologies, the current guideline will be
limited to bile duct lesions only.
It is already known that a significant proportion (70%–80%) of biliary strictures
are malignant, with the risk of malignancy being higher in patients with a definitive
mass on cross-sectional imaging than in those with no clear mass lesion [72]. Early diagnosis is therefore important in directing patients to the proper therapeutic
strategy. While transpapillary brush cytology or forceps biopsy after sphincterotomy
during ERCP (ERCP-TA) have been the standard sampling modality, EUS-guided tissue
acquisition (EUS-TA) has become a useful addition to the diagnostic armamentarium
over time. The pooled sensitivities of brush cytology and intraductal biopsies were
found to be 45% and 48%, respectively, while the combination of the two only modestly
increased the sensitivity to 59% [73]. In contrast, EUS-TA has been reported to have a pooled diagnostic sensitivity of
83% for distal biliary strictures [74]. The appropriate choice of modality will depend not only on the availability of
suitable expertise, but also on the location of the lesion and the presence of jaundice.
Indeed, EUS-TA has shown better sensitivity for distal lesions compared with proximal
lesions [74]. Moreover, in the absence of jaundice, ERCP-TA should be avoided.
3.2.1 Benefits
The task force performed a meta-analysis of eight studies [75]
[76]
[77]
[78]
[79]
[80]
[81]
[82] directly comparing EUS-TA with ERCP-TA for cases requiring additional biliary
drainage for jaundice and the characteristics of these studies are outlined in Table 9s. EUS-TA performed significantly better than ERCP+brushing
(ERCP+brushing vs. EUS-TA: OR 0.31, 95%CI 0.12–0.77; I
2=75%) and with a nonsignificant trend even against ERCP-biopsies
(ERCP-TA vs. EUS-TA: OR 0.66, 95%CI 0.38–1.16; I
2=0%) (Fig. 3s).
EUS-TA is superior owing to the fact that EUS can enable optimal visualization of
the distal bile duct mass or thickening thereby appropriately targeting the lesion.
In fact, bile duct masses as small as 5 mm or a wall thickness of >3 mm have been
shown to be best suited for EUS-TA [75]
[83]. Moreover, a meta-analysis (10 studies, 1162 patients) has shown that there is a
14% incremental benefit of EUS for achieving the diagnosis after nondiagnostic brush
cytology [84]. EUS can aid in tumor staging and access to regional or distant lymph nodes or distant
metastases.
The panel analyzed the diagnostic potential of combining the two endoscopic procedures
(EUS-TA and ERCP-TA) for biliary stricture, based on three studies [75]
[76]
[77]. The diagnostic accuracy for the combined approach was 93%–98% compared with
76%–94% and 55%–81% for EUS-TA only and ERCP-TA only, respectively. The combined approach
was significantly superior (Fig. 4s): EUS-TA+ERCP vs. ERCP-TA
alone, OR 6.2, 95%CI 2.7–13.9; I
2=33%; EUS-TA+ERCP
vs. EUS-TA alone, OR 2.6, 95%CI 1.4–5.1; I
2=0%).
This strategy of combining EUS and ERCP is likely to help in reducing the grey zone
of nondiagnostic or indeterminate biliary strictures, when the diagnostic procedures
have to be further optimized; however, a negative report for either EUS-TA or ERCP-TA
cannot exclude malignancy as both have low post-test probabilities [84].
3.2.2 Harms
While there are no data directly comparing ERCP-TA with EUS-TA for patients without
jaundice, meaning they do not require biliary drainage, we could suggest the use of
only EUS-TA over ERCP-TA, given the better performance of EUS-TA and the adverse events
(AEs) associated with ERCP, which include post-ERCP pancreatitis, cholangitis, duodenal
perforation, and bleeding. Although targeting a thickened stenotic area or mass at
the lower end of the bile duct is technically easier, a theoretical risk of penetrating
a nondrained biliary system leading to bile leak or intracholedochal bleeding should
be considered. Nevertheless, the pooled rate of AEs for EUS-TA for biliary strictures
is 1%, with the majority of AEs being mild and self-limiting [74].
Despite the available data, there are certain caveats to its interpretation with respect
to current practice. While most of the EUS-TAs reported in the above-mentioned studies
consisted of EUS-guided fine-needle aspirations (EUS-FNAs), it is EUS-FNB needles
that are now commonly used worldwide. With the advent of the newer generation of end-cutting
biopsy needles, FNB has been found to be superior to FNA [85] in terms of tissue yield. Therefore, with EUS-FNB, the yield of EUS-TA for distal
biliary strictures is expected to improve further, although we still need more data.
Similarly, needle size and ancillary suction techniques are other areas of contention
[86] that might need answers in order to optimize EUS-TA in distal biliary strictures.
3.2.3 Cost-effectiveness profile
The cost-efficacy of combining EUS-TA and ERCP-TA has not been investigated.
3.2.4 Environmental impact
Gastrointestinal endoscopy is an intensive procedure with significant contributions
to greenhouse gas emissions [87]. Until studies addressing the ecologic impact of endoscopy in diagnosing distal
biliary strictures become available, adherence to guidelines is paramount to assure
the lowest possible environmental impact while maintaining a good quality of the medical
act [88]. In this regard, audit of the procedural diagnostic yield and maintenance of a good
standard of quality in endoscopy also avoids unnecessarily repeating endoscopic procedures
and reduces their environmental impact.
3.2.5 Certainty of the evidence
The risk of bias assessment for each study included in the meta-analysis can be found
in Table 10s. Although based only on retrospective studies, the high quality of the studies, lack
of imprecision/indirectness/inconsistency, and, above all, the high magnitude of the
effect (OR of the combined diagnostic approach >5 vs. ERCP alone and >2 vs. EUS alone)
led to uprating of the level of evidence and the strength of recommendation, as per
GRADE policy (Table 5s).
ESGE suggests performing EUS and ERCP in the same session as this approach does not
increase the risk of adverse events nor impair the effectiveness of the techniques.
Conditional recommendation, very low quality of evidence.
Patients with jaundice due to extrahepatic biliary strictures may require both tissue
sampling and biliary drainage. There are no dedicated studies comparing the outcomes
of the strategy of performing EUS-TA and ERCP separately or in the same session for
isolated biliary strictures; however, evidence may be extrapolated from six studies
(five retrospective and one RCT) that cover a wide range of benign and malignant pathology,
with the majority of lesions being pancreatic masses [89]
[90]
[91]
[92]
[93]
[94].
3.2.6 Benefits
Table 11s reports the baseline characteristics of these six studies [89]
[90]
[91]
[92]
[93]
[94]. On analysis, same-session EUS and ERCP had similar cannulation rates (OR 1.0, 95%CI
0.3–2.9; I
2=11%) compared with separate session procedures (Fig. 5s, part a). Performing both procedures in the same session has the benefit of improved tissue
diagnosis from the combined approach, as mentioned previously.
In the only prospective RCT, 180 patients with suspected distal obstructive jaundice
were randomized into three groups: single-session EUS and ERCP, EUS and ERCP in different
sessions, and ERCP-only groups [91]. Performing the two procedures in the same session required similar procedure times
when compared with separate sessions, while anesthetic and endoscopic complications
were similar among all three groups [91]. Aslanian et al. have pointed out that the same-session approach had lower anesthesia
requirements [89]. These data tend to favor the combined EUS and ERCP approach, preferably in the
same session whenever possible, for reasons of procedural optimization and better
patient comfort.
Moreover, where EUS was performed first, it was possible to avoid ERCP in four cases
in the study by Chu et al. [91]. Therefore, the EUS findings can be regarded more as a roadmap to better understand
the anatomy of the obstructed biliary system and guide the therapeutic endoscopic
procedures. Finally, a more expedited patient evaluation, the avoidance of repeated
sedation [89], a shorter time to chemotherapy [93], shorter hospitalization, and lower costs [95] are potential benefits of the combined procedures.
3.2.7 Harms
On analysis, same-session EUS and ERCP did not produce an increase in AEs (OR 1.2;
95%CI 0.7–2.0; I
2=20%) compared with separate sessions (Fig. 5s, part b).
Our analysis did not find differences in the overall AE rate between single and separate
sessions for EUS-TA and ERCP; however, the included studies mainly referred to pancreatic
masses. For biliary lesions, the risk of AEs may be higher. First, the main pancreatic
duct is rarely stenotic in pure biliary lesions, increasing the risk of post-procedural
pancreatitis. Second, the smaller size of biliary lesions compared with pancreatic
masses and the transbulbar approach that is often used during biliary tumor biopsies
make puncture of the lesion more challenging. These harms seem however to be related
to the type of lesion and may be independent of the sequence/combination of the two
procedures.
Performing EUS-TA and ERCP during the same session may require a longer duration of
sedation for the session, potentially resulting in greater use of general anesthesia
[89]
[93]; however, amounts of anesthetic drugs used in the combined procedure seem to be
lower than in separate sessions [89]
[94] and anesthesia-related AEs do not differ between the two strategies [93]
[94]. Vila et al. reported three desaturations (one requiring orotracheal intubation
and one evolving into aspiration pneumonia) in the same-session EUS and ERCP group,
and one desaturation in the separate sessions group, with no significant differences
found [94].
3.2.8 Cost-effectiveness profile
One retrospective single-center study addressed the cost of combined EUS and ERCP
considering the costs of endoscopic instruments, single-use endoscopic devices, anesthetic
drugs, procedure duration, and hospitalization days [95]. The authors compared the estimated costs of the single-session strategy with the
expected costs of separate sessions, calculating a cost-saving of €510 per patient
in the subgroup of patients who underwent single-session EUS-FNA and ERCP for malignant
lesions, mainly owing to reduced hospitalization time, procedure duration, and professional
fees.
3.2.9 Environmental impact
The coupling of the two procedures in the same session theoretically brings the advantage
of lower energy requirements for sedation, postprocedural monitoring, and hospitalization,
therefore theoretically this recommendation should decrease the environmental footprint
of the endoscopic endeavor in diagnosing distal biliary strictures. However, further
studies are necessary to specifically address this issue.
3.2.10 Certainty of the evidence
The risk of bias assessment for each study can be found in Table
12s. The certainty of evidence for all clinical outcomes in this PICO question
was downgraded owing to the fact that it was mainly noncomparative observational studies
that were included (Table 5s). For the two outcomes of
cannulation rate and AE rate, further downgrading was applied because of imprecision
(wide
CIs crossing 1), which downgraded the quality of evidence to very low.
ESGE suggests performing EUS in patients with suspected biliary stricture and/or unexplained
common bile duct dilatation even in the absence of abnormal laboratory tests.
Conditional recommendation, very low quality of evidence.
With the widespread use of high resolution cross-sectional imaging (CECT, MRI, and
MRCP), the incidental finding of CBD dilatation without symptoms and with normal liver
function tests is being increasingly discovered. A dilated CBD is defined as having
a diameter of ≥7 mm and is commonly encountered in older people, and those with previous
cholecystectomy or chronic narcotic use [96]
[97]
[98]. In the absence of any identifiable etiology on imaging, in clinical practice, these
patients are usually referred for EUS examination; however, there is scarce literature
on the yield of EUS in this setting. Moreover, a further challenging scenario could
be the assessment of the ampulla, where flat lesions can be missed during cross-sectional
imaging and EUS evaluation. Although evidence in the literature is lacking, the use
of side-viewing endoscopes/duodenoscopes for ampullary assessment in patients with
distal biliary stricture without a detectable mass could be a useful addition to the
diagnostic algorithm in this specific setting.
3.2.11 Benefits
The task force performed a meta-analysis, finding seven retrospective studies and
one
previous meta-analysis focusing on the diagnostic value of EUS for this indication
(Table 13s) [99]
[100]
[101]
[102]
[103]
[104]
[105]
[106].
The rate of abnormal findings in these studies was 6% to 35%, with the etiologies
being mostly benign (CBD stones in a large majority) and very rarely malignant (0%–5%
for ampullary and pancreatic tumors). Oppong et al. demonstrated that previous cholecystectomy
is significantly associated with negative EUS findings in the group of patients with
isolated CBD dilatation (P=0.005) [106]. In the study of Kaspy et al., univariate, bivariate, and multivariate logistic
regression analyses demonstrated significant negative associations between an EUS
finding of CBD stones or sludge with history of cholecystectomy (OR 0.23; P=0.03) and age <65 years (OR 0.05; P=0.03), with mean age tending to be higher in patients with CBD stones or sludge (78
vs. 66 years; P=0.07) [103]. In contrast, they found no association between symptoms (biliary-type abdominal
pain) and EUS diagnosis of CBD stone or sludge.
The yield of EUS examination in this setting was evaluated in a systematic review
and meta-analysis by Chhoda et al. that included eight full manuscript studies and
three conference abstracts [99]. Among 224 asymptomatic patients with CBD dilatation, the cumulative yield of EUS
for any pathology was 11% (95%CI 4%–22%) [99]. The EUS yield for benign etiologies was 9% (95%CI 1%–22%), including choledocholithiasis
(3%, 95%CI 0%–11%), whereas for malignant pathology it was 5% (95%CI 0%–3%) [99].
The meta-analysis of the available data showed a rate of abnormal findings of 15%
(95%CI 9%–21%), whereas the rate of malignancy was <0.1% (95%CI 0%–0.1%) (Fig. 6s).
Despite their limitations (retrospective single-center design, limited sample size,
and heterogeneity of the outcome measures), all these studies showed a low, but not
insignificant, rate of benign abnormal conditions (most commonly CBD stones) during
endosonographic exploration of asymptomatic biliary dilatation, demonstrating the
potential benefit of performing EUS in patients with suspected biliary stricture even
in the absence of abnormal laboratory tests. Based on the evidence above, the panel
suggests the use of EUS in this setting, even though neoplastic conditions are extremely
rare.
3.2.12 Harms
No AEs were reported in patients undergoing EUS exploration for suspected biliary
structures with normal laboratory tests. Diagnostic EUS is considered a low risk procedure,
with rare complications such as bacteremia, bleeding, duodenal perforation, and sedation-related
complications, comparable with that of diagnostic upper gastrointestinal endoscopy
[107] and must be factored against the diagnostic yield of EUS, after shared decision-making
with the patient.
3.2.13 Cost-effectiveness profile
Despite its resource use, EUS seems cost-effective for this indication by avoiding
additional imaging examinations when no pathologic finding is detected.
3.2.14 Environmental impact
In a context where the rate of malignancy is <0.1% (as presented above),
justification for a resource-intensive endoscopic procedure that is a generator of
greenhouse gasses and waste is easily debatable. Because appropriateness criteria
are not
perfect, they should always be combined with clinical judgment, with case-by-case
analysis
and decision-making warranted in this context [88].
3.2.15 Certainty of the evidence
The risk of bias assessment for each study can be found in Table
14s. The certainty of evidence for all clinical outcomes in this PICO question
was downgraded owing to the fact that there were only noncomparative observational
studies
included (Table 5s). For the two outcomes of pooled malignancy
rate and rate of abnormal findings, further downgrading was applied because of
inconsistency (high heterogeneity), so downgrading the quality of evidence to very
low.
3.3 Perihilar biliary strictures
ESGE suggests that brushing cytology should be completed along with fluoroscopy-guided
biopsies, wherever technically feasible, in patients with perihilar biliary strictures.
Conditional recommendation, low quality of evidence.
A recent meta-analysis of 21 observational studies [108] concluded that fluoroscopy-guided biopsies plus brush cytology during ERCP for the
evaluation of biliary strictures could lead to a 20% increased diagnostic yield (95%CI
9%–31%) for malignancy versus brush cytology alone. Moreover, the combination of
fluoroscopy-guided biopsies and brush cytology was associated with the highest sensitivity
(66%, 95%CI 63%–69%) compared with brush cytology alone (40%, 95%CI 37%–43%) and
fluoroscopy-guided biopsies alone (52%, 95%CI 49%–56%). Of interest, this analysis
included
both proximal and distal strictures with no difference found between them in terms
of
outcome on subgroup analysis. It should be acknowledged that intraductal fluoroscopy-guided
biopsies are more demanding and technically challenging, asking for advanced expertise.
In
addition, one should keep in mind that intraductal biopsies carry a risk of complications
such as bleeding, duodenal or bile duct perforation, and pancreatitis, and so, whenever
possible, previous multidisciplinary team discussion is recommended in order to decide
on
the optimal approach.
3.3.1 Benefits
The panel did not identify any RCTs evaluating cytology brushing versus fluoroscopy-guided
biopsies, or their combination for proximal biliary strictures. Two cohort studies
[109]
[110] comparing the efficacy of brushing, fluoroscopy-guided biopsies, and their combination
for proximal biliary strictures were retrieved (Table 15s). In the first study, 58 patients with hilar CCA were included and the sensitivities
of brushing cytology, fluoroscopy-guided biopsies, and their combination were 41%,
53%, and 60%, respectively [109]. In the second study, 43 patients with proximal biliary strictures were enrolled
and underwent both cytology brushing and fluoroscopy-guided biopsies using a particular
double-balloon enteroscopy forceps that was advanced to the point of interest through
a plastic pusher after removal of the guidewire [110]. The sensitivity of brushing cytology was 49%, but was 69% for fluoroscopy-guided
biopsies, and increased to 80% when both methods were combined.
The meta-analysis of these two studies showed an incremental diagnostic yield of 25%
(95%CI 11%–38%; P<0.001; I
2=0%) (Fig. 7s, part a) and a significantly increased sensitivity (OR 2.7, 95%CI 1.5–4.9; P<0.001; I
2=0%) (Fig. 7s, part b) for diagnosing malignancy when adding fluoroscopy-guided biopsies to brushings in
perihilar biliary strictures.
Transpapillary biopsies can be performed with standard biopsy forceps. Nevertheless,
studies evaluating slim forceps with a design that allows the angle of the forceps
to be modified have shown good results [111]
[112] that appear to be even better than those obtained with conventional biopsy forceps
[113].
Some reports have shown an increase in diagnostic yield with an increasing number
of biopsies: in a retrospective study, after three or more biopsies, a sensitivity
of 100% was found [114] and, in a prospective study of 31 patients, the reported sensitivity was 55% after
three biopsies, and 59% after four or more biopsies [115]. However, in the most recent meta-analysis available, when comparing the 13 studies
that reported the number of biopsies taken in each study, there was no difference
in sensitivity between studies that performed fewer than four biopsies and studies
that performed four or more biopsies (69% vs. 70%) [116]. Therefore, three to four biopsies may be an adequate number.
Several technical variants have been described to facilitate the ascent of the biopsy
forceps through the bile duct with good results. Although the evidence comes from
small series, given the difficulty of access and lower diagnostic yield in biliary
strictures with more proximal locations, it may be useful to use guidance systems
to reach such strictures [117]
[118]
[119]
[120].
Studies of dilation of the stenosis prior to biopsy are scarce and, although it appears
to increase diagnostic yield without increasing complications, there are insufficient
data to recommend its routine use [121]
[122].
The evaluation of different brushes over several studies has not shown a significant
increase in diagnostic performance that would justify the use of any biliary brushes
other than the conventional ones [123]
[124]
[125].
Recent data from a multicenter RCT with more than 400 included patients showed that
the sensitivity of cytology increased with the number of the brush passes [126]. In this study, the sensitivities were 38%, 47%, and 57% when 10, 20, and 30 brush
passes were performed, respectively. These results encourage increasing the number
of brush passes beyond 10, as this measure would barely increase procedural time and
AEs, and would improve the diagnostic performance of the technique.
3.3.2 Harms
The previously mentioned meta-analysis of studies with both proximal and distal strictures
[108] did not find any significant difference in the rate of AEs between the two groups
(OR 0.53, 95%CI 0.14–2.05); overall, the number of AEs was low, allowing both techniques
to be considered safe to perform. However, two cases with severe AEs including bleeding
and perforation have been described in the setting of intraductal biopsies [127]
[128], so the panel calls for cautious implementation by experienced physicians only,
especially when considering that proximal strictures may be more difficult to approach,
requiring higher technical skills to manipulate and advance the biopsy forceps.
3.3.3 Cost-effectiveness profile
One cost–utility analysis comparing five different sampling techniques among patients
with suspected CCA and PSC did not reveal any significant difference between brushing
and biopsy forceps sampling in terms of incremental QALYs and cost [129]. However, one could reasonably assume that, when both modalities are combined, it
could lead to an increased procedural cost.
3.3.4 Environmental impact
It is known that processing of biopsy/cytology samples obtained during endoscopy leads
to high greenhouse gas emissions; however, this is an unavoidable process. Despite
the fact that combining two sampling modalities (brushing and biopsy forceps) in cases
of proximal biliary strictures would definitely increase the environmental impact
of the procedure as two devices and two specimen pots should be used, one cannot oversee
the potential benefit of avoiding a second ERCP following a positive result after
combining the two methods. Therefore, appropriate patient selection after evaluation
by cross-sectional imaging remains crucial for minimizing the environmental impact
of the procedure [87]
[130].
3.3.5 Certainty of evidence
The risk of bias assessment for each study can be found in Table
16s. The certainty of evidence for all clinical outcomes in this PICO question
was downgraded owing to the fact that there were only observational studies included
(Table 5s). Therefore, the panel concluded by making a
conditional recommendation with a low quality of evidence.
ESGE suggests EUS-TA for perihilar strictures when ERCP-based modalities yield insufficient
results, provided that curative resection is not feasible and/or when cross-sectional
imaging has shown accessible extraluminal disease.
Conditional recommendation, low quality of evidence.
The combination of ERCP-based brush cytology and forceps biopsies provides high
specificity in the diagnosis of perihilar strictures, but may yield insufficient results
in a significant proportion of patients, culminating in a low diagnostic sensitivity.
EUS-TA has proven useful in increasing the diagnostic yield. Although it is regarded
as
safe with low overall complication rates (<1%) and carries a low risk of bile leakage
during EUS-FNA for biliary strictures (2.6%) [131], some specific considerations apply to patients with perihilar disease.
3.3.6 Benefits
A single-blinded comparative study confirmed superior sensitivity and diagnostic
accuracy for EUS-FNA in suspected malignant biliary strictures, compared with ERCP-based
brush cytology + forceps biopsies (94% and 94% vs. 50% and 53%, respectively) [79], results that are in line with more recent studies on the topic [76]
[132]. In the context of perihilar disease in particular, one of the first prospective
studies evaluating the yield of EUS-FNA in surgical patients with negative brush cytology
(n=44) was published in 2004 and showed high diagnostic accuracy (91%) and sensitivity
(89%) [133]. Most importantly, EUS and EUS-FNA changed the surgical approach in almost
two-thirds of patients, without resulting in complications.
When comparing the diagnostic accuracy of EUS-TA in perihilar and distal strictures
however, significant differences in sensitivity and negative predictive value have
surfaced. A 2020 prospective evaluation of EUS-FNA in the diagnosis of biliary strictures
(n=97; 46% hilar) showed an overall sensitivity of 75% (95%CI 64%–84%), with subgroup
sensitivity and negative predictive value reaching 95% and 93% for distal lesions
without stenting, but being much lower (56% and 33%, respectively) in patients with
stented perihilar lesions [134]. While it should be noted that only early generation 22–25G FNA needles were used
in this study, hilar location, underlying PSC, and previous stenting were all identified
as independent risk factors for inaccurate histologic diagnosis using EUS-TA.
Lower EUS-FNA sensitivity in proximal CCA has been reproduced in several studies.
Differences in tumor biology, absence in hypoechogenic mass-like appearance, and diffuse
axial growth patterns may complicate successful identification of the proximal lesion
itself [135]
[136]. Although no comparisons between distal and proximal strictures exist to date, it
has been suggested that contrast-enhanced harmonic EUS might increase T-staging adequacy
[132]
[137]. A previous meta-analysis [138] found a 15% (95%CI 9%–24%) incremental benefit of EUS after inconclusive ERCP with
brushing; however, in this study, patients with both proximal and distal strictures
were included. Moreover, in a recent meta-analysis of 11 studies including patients
with both perihilar and proximal strictures, the pooled sensitivity of ERCP alone
with any tissue acquisition method and EUS alone did not differ (70% [95%CI 66%–73%]
vs. 74% [71%–77%]; P=0.31) [139].
The studies evaluating the role of EUS in the diagnosis of perihilar strictures are
presented in Table 17s. The pooled rates of sensitivity and diagnostic accuracy were 78% (95%CI 70%–86%;
I
2=76%) and 84% (95%CI 78%–91%; I
2=46%), respectively. The task force identified only two comparative studies [76]
[77] that compared EUS-TA and ERCP with brush cytology for patients with perihilar strictures.
The meta-analysis of these two studies showed an incremental diagnostic yield for
adding EUS-TA to ERCP of 33% (95%CI 15%–50%; P<0.001; I
2=0%) (Fig. 8s).
3.3.7 Harms
Peritoneal seeding and subsequent iatrogenic upstaging are theoretical, yet clinically
significant, limitations of EUS-TA and should be weighed against their potential diagnostic
benefit. Even more so, EUS-TA for CCA has been wielded as a contraindication for liver
transplantation [135]. The clear majority of published cases are however limited to needle tract recurrences
in the context of pancreatic adenocarcinoma [140] or following radiofrequency ablation for hepatic lesions [141]. A Japanese nationwide survey suggested an incidence rate of 0.33% in the context
of PDAC, with a significantly higher rate of needle tract recurrences in patients
with PDAC who had undergone transgastric EUS-TA (0.86%) [142]. A pattern that might be explained by the needle tract location lying beyond the
surgical resection margins.
Several retrospective series have evaluated the risk of peritoneal carcinomatosis
following EUS-FNA in the context of PDAC, where it seems to outperform percutaneous
FNA (2% vs. 16%) [143] and no differences were seen when comparing EUS-FNA with alternative methods of
tissue acquisition [144]
[145]. A recent meta-analysis of 10 studies (13238 patients) showed that the pooled rate
of needle tract seeding after EUS-TA of pancreatic lesions was 0.3% (95%CI 0.2%–0.4%)
and no difference was observed in terms of metachronous peritoneal dissemination between
patients who underwent EUS-TA and nonsampled patients (OR 1.0, 95%CI 0.7–1.5; P=0.31), thereby confirming the low risk of seeding with this procedure [146].
Little evidence on EUS-TA-related peritoneal seeding is available in hilar CCA, with
one specifically designed retrospective study (n=150) showing no significant impact
of needle passes on overall or progression-free survival [147]. A comparative study in patients with CCA undergoing transperitoneal FNA before
neoadjuvant chemoradiotherapy and liver transplantation showed higher rates of peritoneal
metastasis (83%) in patients with a positive transperitoneal FNA [148]. The risk of needle tract seeding seems higher however following a percutaneous
approach than with EUS [143], making application of these recommendations to EUS-TA more difficult.
While data on exclusively proximal strictures are lacking, a recent meta-analysis
of studies with both proximal and distal strictures showed that EUS-TA was associated
with significantly fewer AEs compared with ERCP-based sampling (OR 8.1, 95%CI 3.0–22.3),
with minor bleeding being the main EUS-associated AE [139].
3.3.8 Cost-effectiveness profile
While the cost of EUS with FNA/FNB varies across Europe depending on the local market,
ERCP resulted in 9.05 QALYs and a cost of $34685, for a cost-effectiveness ratio of
$3832, while EUS resulted in an incremental increase of 0.13 QALYs and $2773, for
an incremental cost-effectiveness ratio (ICER) of $20840 per QALY gained [40].
3.3.9 Environmental impact
There are no data about the potential environmental footprint of EUS-TA for proximal
strictures; however, the ecologic burden of an additional procedure should be taken
into account (endoscope reprocessing, use of FNA/FNB, sample processing, and other
aspects). Furthermore, as for every endoscopic procedure, the presence of an appropriate
indication to avoid unnecessary examinations should be recorded prior to the examination
[88].
3.3.10 Certainty of evidence
The risk of bias assessment for each study can be found in Table
18s. The certainty of evidence for all clinical outcomes in this PICO question
was downgraded owing to the fact that there were only observational studies included
(Table 5s). Therefore, the panel concluded by making a
conditional recommendation with a low quality of evidence.
3.4 Intraductal Imaging
ESGE suggests using standard ERCP diagnostic modalities at index ERCP. In the case
of indeterminate biliary strictures, ESGE suggests taking cholangioscopy-guided biopsies,
in addition to the standard ERCP diagnostic modalities. Additional intraductal biliary
imaging modalities can be selectively used, based on clinical context, local expertise,
and resource availability.
Conditional recommendation, very low quality of evidence.
The interpretation of real-time images during cholangioscopy may aid in distinguishing
between benign and malignant strictures. Malignant lesions typically appear as either
nodular, papillary, or infiltrating masses. Nodular masses usually lead to luminal
obstruction and exhibit irregular mucosa with pronounced neovascularization. Papillary
masses, on the other hand, feature numerous papillary projections and less neovascularization,
while infiltrating masses cause luminal narrowing without a distinct mass, but display
more whitish mucosal discoloration and neovascularization [149].
The documented sensitivity of visual assessment for identifying malignancy in strictures
varies from 64% to 95% [149]
[150], with most evidence on cholangioscopy being derived from clinical experiences with
single-operator cholangioscopy (SOC). In fact, while other cholangioscopy methods,
such as
direct cholangioscopy and the dual-operator mother–baby technique, may be employed
selectively depending on the clinical scenario, local proficiency, and resource
availability, SOC should be regarded as the preferred technique for cholangioscopy
(Table 19s). In terms of visual assessment performance, image quality
has improved over the last few years, and artificial intelligence-based systems have
been
specifically trained with SOC images [151]
[152], potentially enabling the avoidance of significant interobserver variation, and/or
targeting biopsies. Nevertheless, further study is required to assess the utility
of
artificial intelligence before a recommendation can be issued.
Endoscopic impressions alone are therefore not considered a recognized standard for
making oncologic or other critical decisions. The primary utility of cholangioscopy
lies in tissue acquisition; however, recognizing the distinguishing features of malignant
strictures can facilitate the targeting of cholangioscopy-guided biopsies, potentially
enhancing the diagnostic efficacy of this technique.
Alternative methods for assessing biliary strictures, including intraductal ultrasound
(IDUS) and confocal laser endomicroscopy (CLE), are still evolving and have not established
a concrete role in clinical practice. IDUS findings suggestive of malignancy include
eccentric wall thickening, destruction of wall layers, intraluminal mass with irregular
margins possibly invading surrounding tissue, and heterogeneous lesions with uneven
mucosal surfaces [152]
[153]
[154]. Similarly, CLE employs slender confocal laser probes inserted through the duodenoscope's
working channel, and malignant strictures manifest as thick dark bands of collagen
fibrils and thickened white bands within vessels [155]. Both IDUS and CLE have been shown to improve sensitivity in diagnosing malignant
strictures when compared with ERCP alone; however, these procedures are costly and
require considerable skill to master, with an anticipated low level of interobserver
agreement, thereby impeding their widespread adoption in the near future.
3.4.1 Benefits
Focusing on cholangioscopy, our de novo literature search identified 12 studies (including
one RCT) (Table 20s) assessing the diagnostic performances of standard ERCP diagnostic modalities, including
fluoroscopy-guided biopsies with/without brush cytology, compared with ERCP with cholangioscopy
[80]
[156]
[157]
[158].
The analysis revealed that the incremental yield of cholangioscopy-guided biopsies
was 27% (95%CI 10%–45%), with a significant superiority compared with standard ERCP
(P=0.002; I
2=62%) (Fig. 9s). Furthermore, the sensitivity of ERCP with cholangioscopy was notably higher than
standard ERCP (74% [95%CI 62%–86%] vs. 50% [28%–72%]; OR 1.1, 95%CI 1.0–1.3; I2
=64%) (Fig. 10s). In the case of biliary strictures of undetermined etiology despite previous ERCP
with standard diagnostic modalities (indeterminate biliary strictures), given the
incremental sensitivity yield, patients should be referred to specialized tertiary
centers for cholangioscopy.
3.4.2 Harms
In our meta-analysis, no disparities between cholangioscopy and standard ERCP were
found in terms of risk of AEs (OR 1.5, 95%CI 0.8–2.5; I
2=0%) (Fig. 11s), with pancreatitis and cholangitis being the most frequently reported. In this regard,
the evidence regarding the potential benefits of antibiotic prophylaxis for preventing
post-endoscopy cholangitis in cholangioscopy is limited, retrospective, and conflicting
[159]
[160]
[161]
[162]
[163]. Of note, one study reported that the additional time required for cholangioscopy
during ERCP was around 14 minutes (95%CI 10–20 minutes) [157]. Moreover, access to cholangioscopy remains largely confined to tertiary referral
centers with endoscopists specifically trained in intraductal biliary imaging modalities.
3.4.3 Cost-effectiveness profile
One study detected an additional expense of $2637 when incorporating cholangioscopy
into ERCP [164]; however, it has been shown that, in selected cases, cholangioscopy yields cost-effective
outcomes, reducing the need for multiple procedures and associated expenses in diagnosing
malignancies. In fact, when assessing the economic ramifications of endoscopy procedures
and hospital admissions necessary for the diagnosis and management of indeterminate
biliary strictures, with and without the use of SOC, the implementation of cholangioscopy
demonstrated cost-effectiveness, resulting in a reduction in both the number of procedures
(relative reduction of 31%) and associated costs (decrease of €13000, representing
a relative variation of 5%) when compared with standard ERCP [164].
3.4.4 Environmental impact
There are no data about the potential environmental footprint of cholangioscopy for
biliary strictures. Waste production derived from the use of disposable devices and
endoscopes (i.e. single-use cholangioscopes) is one of the main sources of direct
carbon
emissions in endoscopy [165]. Therefore, combining cholangioscopy with standard ERCP would increase the
environmental impact of the procedure. Moreover, an adjunctive sampling modality
(cholangioscopy-guided biopsies) would translate into further increases in the indirect
sources of emissions derived from processing of histologic samples. On the other hand,
the
incremental sensitivity yield of cholangioscopy compared with standard ERCP may lead
to a
reduction in the number of unnecessary re-examinations, with theoretical benefit in
terms
of avoiding procedure-related sources of direct carbon emission, and indirect sources
such
as unnecessary patient transportation [87].
3.4.5 Certainty of the evidence
The risk of bias assessment for each study can be found in Table
21s. The certainty of evidence for all clinical outcomes in this PICO question
was downgraded because it was based mainly on nonrandomized studies (Table 5s). For the two outcomes of incremental yield and AE rate, further
downgrading was applied owing to indirectness (heterogeneity in the standard ERCP
modalities for tissue sampling), so downgrading the quality of evidence to very
low.
