Introduction
Indeterminate biliary duct stricture (IBDS) frequently poses a diagnostic challenge,
and an accurate diagnosis remains problematic [1 ]. Traditional diagnostic methods such as endoscopic retrograde cholangiopancreatography
(ERCP) with brushing or fluoroscopy-guided intraductal biopsies have a high specificity,
but low sensitivity, for the diagnosis of malignancy [2 ]. A number of benign conditions, such as primary sclerosing cholangitis and IgG4-related
sclerosing cholangitis, frequently mimic malignancy and differentiation can be challenging
[3 ]. Approximately 15 % to 25 % of patients who undergo surgical resection for biliary
strictures because of presumed malignancy, without prior pathological diagnosis, have
benign disease at post-surgical pathological examination [4 ]
[5 ]. Hence, improvement in defining the etiology of a biliary stricture is critical
for improving appropriate curative surgical resection rates for cancer and reducing
the number of unnecessary surgical resections or explorative surgeries.
Digital single-operator cholangioscopy (dSOC) has improved the detection rate of malignancy
through direct stricture visualization and the ability to obtain targeted biopsies
[6 ]
[7 ]. However, the diagnostic accuracy of cholangioscopic visual diagnosis has been shown
to be suboptimal, and therefore obtaining a tissue diagnosis remains mandatory [8 ]. There is limited evidence and currently no consensus on the optimal biopsy strategy
when performing dSOC in patients with IBDS [9 ], with data suggesting that obtaining at least three single biopsies yields more
adequate specimen samples for pathological examination [10 ]. However, improved forceps designs, such as the SpyBite Max forceps (biopsy jaw
outer diameter 1 mm; Boston Scientific, Marlborough, Massachusetts, USA), can obtain
larger tissue samples, but improved diagnostic accuracy has not yet been shown [11 ]
[12 ]. Studies on bite-on-bite biopsies (BBB) for (residual) esophageal cancer [13 ] and subepithelial lesions [14 ]
[15 ] have suggested a superior yield for establishing a definitive pathology diagnosis
compared with single biopsies, albeit with a large range. It has been suggested that
instead of obtaining single biopsies, BBB may obtain larger and deeper bile duct samples
and have a higher diagnostic yield for diagnosis of malignancy in patients with IBDS.
To date, no studies have systematically compared the diagnostic yield of four single
biopsies with that of BBB in patients with IBDS. We aimed to assess the feasibility
and diagnostic yield of tissue sampling via multiple dSOC-guided biopsies and to analyze
the added yield of the bite-on-bite (-on-bite) technique in a representative study
population with a high pretest probability of malignancy.
Methods
Study design and population
An international, multicenter, prospective interventional study was performed in patients
who underwent dSOC for the evaluation of IBDS between November 2020 and August 2022.
All 14 participating centers were members of the European Cholangioscopy Group, including
therapeutic endoscopists with significant experience of dSOC (> 200 lifetime procedures).
All centers obtained ethical approval from their respective local ethics committees.
The study was conducted according to the principles of the Helsinki Declaration, and
followed the Strengthening the Reporting of Observational studies in Epidemiology
(STROBE) reporting guidelines (see Table 1 s in the online-only Supplementary material).
Eligible patients were diagnosed with an IBDS, defined as a bile duct stricture or
filling defect of indeterminate nature after previous laboratory work-up, abdominal
imaging, ERCP or endoscopic ultrasound (EUS), and/or negative or inconclusive cytology
or histology after ERCP with brush cytology or fluoroscopy-guided intraductal biopsies,
and/or despite indefinite cytology or histology findings with persistent suspicion
of a malignant etiology based on clinical symptoms such as unexpected weight loss,
loss of appetite, and obstructive jaundice. Other inclusion criteria were 18 years
of age or older, and willingness and ability to provide written informed consent for
participation. Patients were not eligible when a contraindication to ERCP, dSOC, or
biopsy was present.
dSOC procedure
Procedures were performed with the patient under conscious sedation, propofol sedation,
or general anesthesia. Cholangioscopy was performed with a digital single-operator
cholangiopancreatoscopy system (SpyGlass DS II Direct Visualization System; Boston
Scientific), which was advanced either over a 0.025– or 0.035-inch guidewire, or freehand
into the common bile duct (CBD) up to the level of the target stricture. Examples
of IBDS under dSOC view are presented in [Fig. 1 ]. After successful identification of the stricture in each patient, at least four
single visually guided biopsies were obtained, followed by at least one BBB performed
with the Spybite Max Forceps.
Fig. 1 Three examples of indeterminate biliary duct strictures with digital single-operator
cholangioscopy. a A perihilar stricture. Multiple single biopsies showed benign disease, while bite-on-bite
biopsy (BBB) yielded inadequate tissue. Finally, endoscopic ultrasound-guided biopsy
confirmed the suspicion of perihilar cholangiocarcinoma. b A distal stricture located at the insertion of the cystic duct. Both single and BBB
biopsies were inconclusive. Surgical resection confirmed chronic segmental sclerosing
cholangitis. c A perihilar stricture. Both single and BBB biopsies were suspicious for malignancy.
All endoscopists had the same step-by-step procedure form, ensuring similar biopsy
techniques. Rapid on-site evaluation was not used. There was no centralization of
histopathological reading; all histopathological examinations of biopsies were performed
by an unblinded expert gastrointestinal pathologist at each participating center.
Prophylactic measures for cholangitis and post-ERCP pancreatitis were used according
to local protocols.
During the study, the biopsy protocol was adapted from bite-on-bite (double BBB) to
bite-on-bite-on-bite biopsy (triple BBB) to attempt even deeper biopsies of the bile
duct lesion, as initial observations did not show much promise. A set of BBB was defined
as two samples for double BBB and three samples for triple BBB. Single biopsy refers
to one biopsy taken from one single location, and was attempted at least four times.
Examples of both biopsy techniques are shown in [Video 1 ].
Video 1 Cholangioscopy-guided single biopsy and bite-on-bite biopsies in indeterminate biliary
duct strictures.
Data collection
Baseline data were collected and included patient demographics (age, sex), history
of primary sclerosing cholangitis, IgG4-related sclerosing cholangitis, chronic pancreatitis,
gallstone disease, malignancy, hepatobiliary surgery, previous placement of biliary
stents, and bile duct stone removal, as well as previous diagnostic work-up regarding
the IBDS. In addition, data from previous investigations, such as ERCP and EUS with
tissue acquisition, computed tomography (CT), magnetic resonance imaging (MRI), magnetic
resonance cholangiopancreatography (MRCP), and CA19.9 levels were collected.
During the dSOC procedure, the following variables were collected: technical success
of biopsy sampling, the number of biopsies collected by each technique, and any procedural
adverse events. IBDS site was divided into proximal (intrahepatic, hilar, proximal
CBD, and cystic duct) and distal (distal CBD). Biopsies were evaluated and categorized
by an expert pathologist into the following categories: benign, malignant, or various
degrees of dysplasia (low grade or high grade), with an additional category for inconclusive
results. Inconclusive results were specimens in which the pathologist was unable to
interpret the tissue.
A diagnosis of malignancy could be established by dSOC-guided biopsies, other tissue-sampling
methods during follow-up, surgical resection specimens, or disease progression (i. e.
development of unequivocal distant metastasis) observed on cross-sectional imaging.
A definitive diagnosis of a benign condition was established after a minimum follow-up
period of 1 year, during which there was no confirmation of malignancy using any of
the aforementioned techniques and no documented disease progression on cross-sectional
imaging.
Outcome measures
The primary outcome was the accuracy of both biopsy techniques (i. e. single biopsies
and BBB) for diagnosis of malignancy in IBDS. In addition, the added diagnostic yield
of performing additional BBB biopsies after single biopsies was analyzed. In the statistical
analysis, specimens interpreted as low grade dysplasia and inconclusive diagnosis
were considered inconclusive for malignancy; specimens interpreted as high grade dysplasia
were considered positive for malignancy.
Secondary outcomes were technical success rates for obtaining tissue with both techniques
and the adverse event rate of both techniques. Adverse events related to BBB, consisting
of bile duct perforation and hemobilia, were graded according to the AGREE classification
[16 ]. Other adverse events related to the ERCP or dSOC were noted separately, such as
post-ERCP pancreatitis or cholangitis.
Sample size calculation
Sample size was calculated using a two-sided McNemar’s test for paired two-sample
proportion. At a significance level (α) of 0.05, 76 patients were needed to achieve
the power of 0.9 (β), assuming a net benefit of 15 % of BBB compared with single biopsies.
Data and statistical analysis
Descriptive statistical variables consisted of frequencies (%), medians, and interquartile
ranges (IQRs). Sensitivity, specificity, accuracy, negative predictive value (NPV),
and positive predictive value (PPV) of both biopsy techniques were calculated using
2 × 2 tables according to the final pathological diagnosis. The 95 % CIs were calculated
according to the Clopper–Pearson interval. In this analysis patients who were lost
to follow-up were excluded. The exact McNemar’s test was performed to compare outcomes
among patients in whom both techniques were successful and final pathological diagnosis
was available. A P value of < 0.05 was interpreted as statistically significant. All analyses were performed
using R version 4.2.2. (R Foundation for Statistical Computing, Vienna, Austria).
Results
Baseline characteristics
A total of 89 patients with a new diagnosis of IBDS were prospectively included. The
median patient age was 66 years (IQR: 55–73), and the majority of patients were male
(61.8 %). The IBDS was located proximally in 76 patients (85.4 %) and distally in
13 (14.6 %). Only six patients (6.7 %) had a diagnosis of primary sclerosing cholangitis
prior to dSOC. Prior work-up consisted of CT in 86 (96.6 %), MRI/MRCP in 60 (67.4 %),
EUS in 21 (23.6 %) with fine-needle aspiration/biopsy (FNA/FNB) issue acquisition
in 14 (66.7 %), and ERCP in 41 (46.1 %), with negative or inconclusive brushings/biopsies
in all. In nine patients (10.1%), both an ERCP with tissue sampling and an EUS with
FNA/FNB were performed. Overall, in 46 patients (51.7 %) prior tissue acquisition
through EUS and/or ERCP was performed. All baseline characteristics are listed in
[Table 1 ].
Table 1
Baseline characteristics of the included patients.
Patients (n = 89)
Age, median (IQR), years
66 (55–73)
Female sex, n (%)
34 (38.2)
Stricture location, n (%)
13 (14.6)
4 (4.5)
20 (22.5)
38 (42.7)
14 (15.7)
CA19.9 level, median (IQR), mmol/L[1 ]
74.4 (14.0–294.0)
Medical history, n (%)
Primary sclerosing cholangitis
6 (6.7)
Chronic pancreatitis
4 (4.5)
Choledocholithiasis
10 (11.2)
Hepatobiliary surgery
10 (11.2)
7 (70.0)
2 (20.0)
1 (10.0)
Biliary interventions prior to dSOC, n (%)
Prior stent placement
37 (41.6)
Prior stone removal
7 (7.9)
Work-up of IBDS, n (%)
CT
86 (96.6)
MRI/MRCP
60 (67.4)
EUS
21 (23.6)
14 (66.7)
4 (28.6)
10 (71.4)
ERCP
41 (46.1)
29 (70.7)
15 (51.7)
12 (41.4)
2 (6.9)
1 (2.4)
1 (100)
11 (26.8)
3 (27.3)
8 (72.7)
46 (51.7)
Final pathology outcome, n (%)[3 ]
36 (40.5)
4 (4.5)
1 (1.1)
1 (1.1)
1 (1.1)
6 (6.7)
2 (2.3)
36 (40.5)
2 (2.3)
CBD, common bile duct; CCA, cholangiocarcinoma; CT, computed tomography; dCCA, distal
cholangiocarcinoma; dSOC, digital single operator cholangioscopy; ERCP, endoscopic
retrograde cholangiopancreatography; EUS, endoscopic ultrasound; FNA/FNB, fine-needle
aspiration/fine-needle biopsy; GBC, gallbladder cancer; HCC, hepatocellular carcinoma;
IBDS, indeterminate biliary duct stricture; iCCA, intrahepatic cholangiocarcinoma;
IQR, interquartile range; MRI/MRCP, magnetic resonance imaging/magnetic resonance
cholangiopancreatography; pCCA, perihilar cholangiocarcinoma.
1 Measured in 69 patients.
2 Combination of ERCP and EUS.
3 Based on dSOC in 31, surgery in 18, clinical follow-up in 30, and pathological proof
in 8 patients.
Procedure characteristics
The dSOC procedure was technically successful, with visualization of the target stricture
in all patients. Overall, obtaining single biopsies was technically successful in
88 /89 patients (98.9 %). In one patient, biopsies could not be obtained due to an
inadequate cholangioscope position. A total of 83 patients (93.3 %) had at least four
biopsies taken, and five patients (5.6%) had either two or three biopsies, which did
not conform to the protocol. In 82 /88 patients (93.2 %), the pathologist concluded
that sufficient tissue had been obtained for histopathological analysis.
The BBB technique was considered technically successful in 86 /89 patients (96.6 %).
In three patients, biopsies could not be obtained due to inadequate cholangioscope
position caused by the scope being pushed away during BBB. A median number of two
sets of BBB (IQR: 2–3) were obtained. In 66 patients (76.7 %), two or more sets of
BBB were obtained. In 20 patients (23.3 %) double BBBs were obtained, while in the
remaining 66 patients (76.7 %) triple BBBs were obtained. Adequate tissue samples
for histopathological analysis were obtained with BBB in 78 patients (87.6 %), as
determined by the pathologist. No adverse events directly associated with BBB, such
as hemobilia or bile duct perforation, were noted. Mild post-ERCP pancreatitis developed
after dSOC in two patients, both of whom were treated conservatively (Grade II). One
patient had cholangitis treated conservatively with antibiotics (Grade II), while
another patient with cholangitis underwent stent exchange (Grade IIIa).
Biopsy technique outcomes
Among the 82 patients with sufficient tissue from single biopsies available for analysis,
histopathological reports revealed malignancy in 31 patients (37.8 %), high grade
dysplasia in 2 patients (2.4 %), low grade dysplasia in 2 patients (2.4 %), benign
disease in 38 patients (46.3 %), and inconclusive in 9 patients (11.0 %).
Among the 78 patients with sufficient tissue from BBB available for analysis, histopathological
reports revealed malignancy in 29 patients (37.2 %), high grade dysplasia in 1 patient
(1.3 %), low grade dysplasia in three patients (3.8 %), benign disease in 37 patients
(47.4 %), and inconclusive in 8 patients (10.3 %).
In 76 patients (85.4 %), both single biopsies and BBB were available and sufficient
for analysis, as shown in [Fig. 2 ]. In 73/76 patients (96.1 %) histopathological diagnosis was similar for both techniques.
Three patients (3.9 %) had discordant results between the two techniques: 1) in one
patient, single biopsies confirmed malignancy consistent with perihilar cholangiocarcinoma,
but BBB showed benign disease; 2) in another patient, single biopsies were inconclusive,
but BBB confirmed benign disease in line with clinical follow-up at > 1 year; and
3) in the remaining patient, single biopsies showed benign disease, and BBB showed
low grade dysplasia, while surgical resection after 7 months identified benign sclerosing
cholangitis.
Fig. 2 Flow chart of included patients. *All five patients with < 4 single biopsies had
adequate specimens.
BBB showed an additional yield compared with single biopsies in only 3/89 patients
(3.4 %). In the abovementioned patient with inconclusive single biopsies, BBB confirmed
benign disease, and in two patients in whom single biopsies were insufficient for
histopathological assessment, BBB correctly confirmed benign disease.
Final diagnosis after dSOC and follow-up
Overall, dSOC-guided biopsies confirmed malignancy in 31/89 patients (34.8 %) and
high grade dysplasia in 2/89 patients (2.2 %). Among the remaining 53 patients with
inconclusive, benign, or low grade dysplasia on histopathological analysis from biopsies,
malignancy was confirmed after the dSOC procedure in 12/53 patients (22.6 %). This
diagnosis was based on surgery in eight patients, additional ERCP-guided brushings
or biopsies in two patients, additional EUS-guided tissue acquisition in three patients,
and evident tumor progression on cross-sectional imaging in two patients. In the three
patients with low grade dysplasia, surgical specimens confirmed benign sclerosing
cholangitis in one, repeat ERCP-guided brushings showed distal cholangiocarcinoma
in another, and EUS-guided tissue acquisition of the mass showed perihilar cholangiocarcinoma
in the remaining patient. Regarding the two patients with high grade dysplasia in
dSOC-guided biopsies, one patient had ERCP-guided biopsies with proven intrahepatic
cholangiocarcinoma after 5.4 months, and the other patient had evidence of tumor progression
on cross-sectional imaging after 5.1 months. In 36/56 patients (64.3 %), benign disease
was confirmed through surgical resection in 9 patients and clinical follow-up in 27,
with a median follow-up period of 12 months (IQR: 8–17). Two patients were lost to
follow-up.
Diagnostic accuracy of both biopsy techniques
Overall sensitivity, specificity, PPV, NPV, and accuracy for diagnosing malignancy/high
grade dysplasia in IBDS were 66.0 %, 100 %, 100 %, 63.8 %, and 78.8 %, respectively,
for single biopsies, and 63.8 %, 100 %, 100 %, 63.0 %, and 77.6 %, respectively, for
BBB ([Table 2 ]). Accuracy was not significantly different between biopsy techniques (P > 0.99). Whenever a previous ERCP with tissue acquisition was performed prior to
dSOC, for both single biopsies and BBB, sensitivity, NPV, and overall accuracy were
lower, as shown in [Table 3 ]. These results were comparable to subgroup analysis with or without prior stent
placement. Results regarding double or triple BBB are presented in Table 2 s and show a slightly higher accuracy for double BBB.
Table 2
Outcomes of both biopsy techniques.
Bite-on-bite (-on-bite)
Single biopsies
Technical success, n (%)
86 (96.6)
88 (98.9)
Number of biopsies, median (IQR)
2 (2–3)
4 (4–4)
Number of biopsies, n (%)
20 (23.3)
0
30 (34.9)
1 (1.1)
24 (27.9)
4 (4.5)
12 (14.0)
73 (83.0)
0
7 (8.0)
0
3 (3.4)
Double or triple bite, n (%)
20 (23.3)
NA
66 (76.7)
NA
Adequate tissue, n (%)
78 (87.6)
82 (92.1)
Pathology outcome of biopsy, n (%)
37 (47.4)
38 (46.3)
8 (10.3)
9 (11.0)
3 (3.8)
2 (2.4)
1 (1.3)
2 (2.4)
29 (37.2)
31 (37.8)
Final pathology results available
76[1 ]
80*
30 (39.5)[2 ]
33 (41.3)[3 ]
17 (22.4)[4 ]
17 (21.3)[4 ]
0
0
29 (38.2)
30 (37.5)[5 ]
63.8 (48.5–77.3)
66.0 (51.2–78.8)
100 (88.1–100)
100 (88.4–100)
100 (88.4–100)
100 (89.4–100)
63.0 (53.9–71.4)
63.8 (54.5–72.2)
77.6 (66.6–86.4)
78.8 (68.2–87.1)
CCA, cholangiocarcinoma; iCCA, intrahepatic cholangiocarcinoma; IQR, interquartile
range; NA, not assessable; pCCA, perihilar cholangiocarcinoma.
1 Final pathology results were missing in two patients.
2 Including one patient with high grade dysplasia and confirmed pCCA.
3 Including two patients with high grade dysplasia and confirmed CCA (1 pCCA, 1 iCCA).
4 Including six patients with inconclusive biopsies and two patients with low grade
dysplasia.
5 Including one patient with no pathological proof after 370 days but moderate suspicion
for pCCA.
Table 3
Subgroup analyses for patients with and without prior endoscopic retrograde cholangiopancreatography
with tissue acquisition.
Bite-on-bite (-on-bite)
Single biopsies
Prior ERCP with tissue (n = 32)
No prior ERCP (n = 44)
Prior ERCP with tissue (n = 37)
No Prior ERCP (n = 43)
True positive, n (%)
9 (28.1)
21 (47.7)
10 (27.0)
23 (53.5)
False negative, n (%)
10 (31.3)
7 (15.9)
11 (29.7)
6 (14.0)
False positive, n (%)
0
0
0
0
True negative, n (%)
13 (40.6)
16 (36.4)
16 (43.2)
14 (32.6)
Sensitivity (95 % CI), %
47.4 (24.5–71.1)
75.0 (55.1–89.3)
47.6 (25.7–70.2)
79.3 (60.3–92.0)
Specificity (95 % CI), %
100 (75.3–100)
100 (79.4–100)
100 (79.4–100)
100 (76.8–100)
PPV (95 % CI), %
100 (66.4–100)
100 (83.9–100)
100 (69.2–100)
100 (85.2–100)
NPV (95 % CI), %
56.5 (45.9–66.6)
69.6 (54.6–81.3)
59.3 (49.2–68.6)
70.0 (53.4–82.6)
Accuracy (95 % CI), %
68.8 (50.0–83.9)
84.1 (69.9–93.4)
70.3 (53.0–84.1)
86.1 (72.1–94.7)
ERCP, endoscopic retrograde cholangiopancreatography; NPV, negative predictive value;
PPV, positive predictive value.
Discussion
This study aimed to assess and ultimately improve the current biopsy techniques for
diagnosing IBDS with dSOC. Although general recommendations include taking multiple
single biopsies from IBDS, deeper sampling capability might offer improved diagnostic
accuracy over standard single biopsies. By correct identification of the nature of
the IBDS, clinical decision making and patient care are influenced. Our findings suggest
that performing BBB biopsies does not increase the diagnostic accuracy in patients
with IBDS.
We found that by adding BBB to the current clinical practice of single biopsies, an
additional yield was obtained in only 3.4 % of patients. In all of these patients,
BBB correctly confirmed benign disease where single biopsies were inconclusive or
inadequate for assessment. Direct comparison of the two techniques showed a sensitivity
and specificity of 66.0 % and 100 %, respectively, for targeted single biopsies, and
63.8 % and 100 % for ≥ 1 set of BBB biopsies. Overall, these rates are in line with
other reports on the yield of dSOC-guided biopsies of biliary strictures [6 ]
[17 ]. However, we have shown that taking one set of deeper samples by dSOC-guided BBB
does not increase the yield. The rationale for taking deeper biopsies is that cholangiocarcinoma
has a prominent desmoplastic and vascularized stroma. Cholangiocarcinoma has three
main patterns of growth: mass-forming, periductal infiltrating, and intraductal growing
[18 ]. In particular, periductal infiltrating cholangiocarcinoma, which may not grow into
the bile duct lumen, is notoriously hard to diagnose with superficial brushes or biopsies.
We discovered that BBB is more difficult to perform successfully than single biopsies,
especially in patients with distal strictures, owing to scope instability and difficulty
identifying the site of the first bite for BBB. Therefore, diagnostic accuracy rates
found might be lower due to sampling difficulty.
It is important to note that there are many factors influencing the yield of dSOC-guided
biopsies, such as the number of patients with a high pretest odds of malignancy. We
performed a subanalysis in patients who underwent prior intraductal tissue acquisition
through ERCP by biliary brushes and/or biopsies. For both single and BBB biopsies,
the sensitivity, NPV, and accuracy were higher in patients with “naïve” IBDS ([Table 3 ]). This is probably due to the selection of patients with IBDS because it may be
harder to acquire diagnostic tissue from patients with previous sampling than in naïve
IBDS without prior ERCP sampling. Central to this discussion is the definition of
a “true” or “naïve” IBDS. While the study was not designed to evaluate dSOC as the
primary method for obtaining tissue in IBDS, the high diagnostic accuracy and sensitivity
of dSOC in “naïve” IBDS cases suggest its potential value. This promising data could
justify further investigation of dSOC as an initial approach in index ERCP for IBDS.
Future research should, however, consider the cost-effectiveness of this method. On
the other hand, recent studies on optimized protocols for biliary brush cytology have
shown a steep increase in sensitivity to around 78 % for perihilar and intrahepatic
cholangiocarcinoma [19 ].
Finally, it is important to note that historically, IBDS has been defined as biliary
stricture in which prior ERCP rendered inconclusive results. Recent Asian consensus
guidelines on the role of dSOC defined an IBDS to be of uncertain etiology with inconclusive/negative
imaging or tissue diagnosis [1 ]. This definition differs critically from the recent American guidelines, which define
IBDS as “one for which a diagnosis has not been established despite initial ERCP with
intraductal sampling” [20 ]. It is important to make the distinction between undetermined and indeterminate
biliary duct strictures and therefore we advocate for a uniform definition internationally,
in order to decrease the heterogeneity of future studies and facilitate easier comparison
[21 ].
Although the Spybite Max is capable of obtaining larger tissue samples, previous studies
have not indicated an increase in diagnostic accuracy [11 ]
[12 ]. Nonetheless, the biggest advancement in reaching definitive diagnosis will likely
come from the possibility of obtaining larger tissue samples, and more studies are
needed on larger biopsy forceps. Other methods of dSOC-guided tissue acquisition include
snares that can remove polypoidal lesions and bile duct aspiration, aimed at circulating
tumor and cell-free DNA analysis. It is likely that in the future additional methods
will be developed, such as dSOC-guided brushes or dSOC-guided FNB. Meanwhile, there
are many promising applications that can complement dSOC-guided biopsies, which were
not taken into account in this study, such as the application of next-generation sequencing
on biopsies [22 ]. Next-generation sequencing may provide additional information by identifying several
genetic mutations associated with cholangiocarcinoma. The introduction of artificial
intelligence in dSOC may guide targeted biopsies, potentially increasing the yield
even further [23 ]
[24 ]. Another promising technique is cryobiopsy, which is designed to overcome several
disadvantages of dSOC-guided biopsies, but has only been performed through the percutaneous
approach [25 ]
[26 ]. Recently, endoscopic light-scattering spectroscopy guided by dSOC showed 97 % accuracy
for detecting malignancy without tissue acquisition [27 ].
A significant strength of our study is the prospective design in which patients underwent
both biopsy techniques to make direct intrapatient comparison possible. We included
both patients with a negative work-up after ERCP or EUS and patients who had not undergone
work-up, allowing us to directly compare both groups with “true” IBDS (after an initially
negative work-up) and casus pro diagnosi (without initial work-up). A structured follow-up
for patients with suspected benign disease after dSOC limited the possibility of false-negative
diagnosis. There are also limitations that need to be considered. First, the method
of histopathological analysis, as well as the pathologists’ experience with IBDS,
were not standardized. This could be improved by centralization of reading by one
dedicated pathologist, but this may limit generalizability to clinical practice. Second,
this study was carried out in expert treatment centers by experienced endoscopists,
limiting the generalizability to lower volume centers while acknowledging that obtaining
BBB biopsies is technically challenging. Third, several characteristics of the IBDS
itself, such as length of the stricture and visual characteristics during dSOC, were
not collected systematically. Fourth, we included high grade dysplasia as a positive
outcome together with malignancy, because high grade dysplasia is treated as being
malignant in the participating treatment centers. Finally, dSOC is an expensive procedure
and there is still limited information pertaining to its cost-effectiveness and its
environmental footprint. Finally, we did not differentiate between the various types
of cholangiocarcinoma. It is conceivable that there is a differential yield between
exophytic and non-exophytic types.
In conclusion, obtaining a set of biopsies through BBB did not produce statistically
significant differences in diagnostic yield compared with the current strategy of
multiple single biopsies. We cannot conclusively refute the hypothesis of equivalent
accuracy between the two strategies. An interesting approach to assessing the value
of BBB would be to collect all consecutive “bites” of the BBB in separate containers
to determine the true added value and possibly the optimal number of consecutive biopsies
at one site in IBDS. Therefore, further studies with larger sample sizes are needed,
assessing the successive yield of BBB per set, for multiple sets, and stratified by
cholangiocarcinoma type, including the addition of next-generation sequencing and
a larger biopsy forceps.