Introduction
Biliary stricture is an abnormal narrowing of the bile duct that can result from a
wide spectrum of benign causes, such as choledocholithiasis, primary sclerosing cholangitis,
abdominal radiation treatment, chronic pancreatitis, traumatic or ischemic injury,
and postsurgical strictures. Yet, approximately 70 % of biliary strictures are neoplastic
in origin, such as resulting from cholangiocarcinoma, pancreatic cancer, gallbladder
cancer, hepatocellular cancer and metastatic carcinomas [1 ]
[2 ]. It is important to accurately characterize the pathology of the biliary stricture
because it is the basis for choosing the appropriate treatment: while endoscopic procedures
may be sufficient for benign diseases, surgery is generally required for treatment
of malignant diseases. However, biliary strictures frequently present a diagnostic
challenge in the clinics.
To date, tissue or cytology acquisition by endoscopic retrograde cholangiography (ERC)
or by endoscopic ultrasonography (EUS) remains the gold standard for diagnosis of
malignant strictures [3 ]. These endoscopy-based methods include biopsy or brush cytology with ERC, biopsy
with choledochoscopy [4 ] and endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) [5 ]. In particular, ERC is the primary and the most widely used endoscopic procedure
for evaluating bile duct strictures, because it provides the opportunities to identify
the biliary stricture, obtain specimens for cytological or histological evaluation,
and perform biliary drainage. Fluoroscopic biopsy and biliary brush cytology are two
major ERC techniques used in transpapillary biliary sampling for the evaluation of
patients with biliary obstruction. For both methods, the specificity for diagnosis
of neoplastic biliary strictures is high but the sensitivity remains unsatisfactory
[1 ]. In recent years, a variety of new biopsy forceps and new methods of forceps biopsy
have been developed to improve sensitivity, such as biopsies under the guidance of
pusher [6 ], ”ropeway-type forceps” with a side slit for a guidewire [7 ], and 90° adjustable biopsy forceps [8 ]. However, these efforts have not led to significant improvement in the yield of
neoplasia.
We have developed fluoroscopy-guided, shaped endobiliary biopsy (FSEB), a unique biopsy
approach in which the biopsy forceps are modified by the physician to mimic the shape
of the bile duct involved. Using the “shaped” biopsy forceps has the benefit of improving
access to the targeted tissue and, thereby, increasing the yield and accuracy of diagnosis.
Here, we demonstrate the diagnostic characteristics of FSEB for neoplastic biliary
lesions in a large cohort of consecutive patients with biliary stricture undergoing
ERC. Our results indicate that FSEB has high sensitivity and accuracy and can serve
as a first-line sampling tool for the evaluation of biliary strictures.
Patients and methods
Patients
The study included consecutive patients who had biliary stricture on the first endoscopic
retrograde cholangiography (ERC), with endobiliary biopsies obtained using FSEB by
the senior author (S.K. Lo) at the Cedars-Sinai Medical Center between January 2001
and December 2014. Patients that underwent choledochoscopic biopsies only and patients
with incomplete medical records were excluded from this study ([Fig. 1 ]). For intention-to-treat analysis, patients were included as false negative if they
dropped out in the follow-up. Patients were excluded if death occurred from unrelated
reasons or if patients chose to terminate follow-up due to advanced age or severe
co-morbidities. The study was approved by the Cedars-Sinai Institutional Review Board
(IRB protocol #23115). The data in this study are reported in accordance with the
STARD guidelines [9 ].
Fig. 1 Flowchart summarizing patient enrollment and study outcomes. The 10 patients who dropped
out before 1-year follow-up were considered false-negative by the strict definition
and excluded in the per protocol analysis.
Location of bile duct stricture and classification of malignant bile duct stricture
Proximal bile duct is confined to the area between intrahepatic bile duct and the
level of the cystic duct. Distal extrahepatic bile duct refers to the area between
the origin of the cystic duct and the ampulla of Vater. The primary malignant bile
duct stricture refers to the biliary stricture induced by the ingrowth of cholangiocarcinoma,
secondary malignant bile duct stricture refers to the biliary stricture caused by
the ingrowth or external compression by pancreatic, gallbladder, liver, duodenal,
or metastatic cancer, lymphomas, or the lymph node.
Endoscopic procedure
The biliary stricture was evaluated by ERC using a standard duodenoscope available
at the time of endoscopy (JF-260V; Olympus Optical, Tokyo, Japan). The instruments
used for deep cannulation of the biliary tree were based on the endoscopist’s choice
based on the clinical and anatomical considerations, such as the angle of entry and
the general direction of the bile duct. A biliary sphincterotomy of native papilla
was generally performed to facilitate further therapy. The stricture was dilated with
a balloon dilator up to either 6 or 8 mm, depending on the location of the stricture
and overall caliber of the bile ducts. Subsequently, FSEB was performed on all patients
in this study.
Creation of FSEB
FSEB is achieved by manually bending (“shaping”) a standard upper endoscopy forceps
(FB-220U; with cup capacity of 6.1 mm3 and jaw opening width of 7.1 mm, Olympus Medical Systems, Tokyo, Japan) to facilitate
negotiation through a duodenoscope elevator and to achieve the angulation needed to
obtain a biopsy ([Fig. 2 ]). The endoscope used in this study has a 4.2-mm working channel and can accommodate
both a guidewire and a biopsy forceps alongside the wire.
Fig. 2 Modification of biopsy forceps. a Images of standard and shaped biopsy forceps. b The orientation of biopsy forceps can be controlled by manually changing wire tension
(selected frames from [Video 1 ]; tension decreased from the left to right panels).
For cases that required proximal biopsies, additional angles were created. The first
angle is made by manually grooming the tip of the biopsy forceps and this angle can
be altered based on the relative position of the duodenoscope to the papilla. The
length of the bile duct is estimated by visual inspection or standard measurement
technique using a cannula. Based on the distance between the tumor and the ampulla,
a second shape/bend is created proximally (away from the forceps) as illustrated in
[Fig. 2 ]. Please see [Video 1 ] for further explanation of creation and use of FSEB.
Video 1 Demostration of orientation of the biopsy forceps by manually changing wire tension.
Biopsy technique
Endoscopic sphincterotomy was performed in all patients prior to FSEB. We have not
attempted passage of these forceps without sphincterotomy for perceived high risk
of ERCP-related pancreatitis. The shaped forceps were introduced into the biliary
duct under endoscopic and fluoroscopic guidance. After introduction into the bile
duct, changing the angle of the tip of the biopsy instrument can be achieved by altering
the pressure of closure ([Fig. 2 ], [Fig. 3 ], and [Video 1 ]). At normal pressure of closure of the forceps, the tip of the shaped forceps may
contain one or two angulations based on the desired “shaping” that was performed by
the endoscopist. Pressure of closure of the forceps handle is increased in order to
straighten the forceps. Occasionally, the forceps can move in a different direction
than desired. In that instance, the biopsy forceps may need to be reshaped based on
the length and direction of the bile duct. Rarely, a new biopsy forceps may be needed
if reshaping the existing forceps is not adequate. Our routine standard practice was
to obtain at least three pieces of tissue for histopathology. If adequate tissue was
obtained, brushing was not performed.
Fig. 3 Images of FSEB for biliary stricture at various sites. a, b Biopsy of hilar stricture. c, d Right and left intrahepatic stricture. e, f Distal bile duct stricture.
Thus, by using this dynamic alteration in pressure of closure of the forceps, the
distal tip of the forceps is carefully negotiated to the desired location for a successful
endobiliary biopsy ([Fig. 2 ], [Fig. 3 ] and [Video 1 ]). Generally, at least three specimens were obtained from different areas of the
stricture, under fluoroscopic observation to guarantee the adequacy of tissue for
histopathological evaluation. The biopsy specimens were fixed in 10 % formalin for
histopathological evaluation. After biopsies were performed, one or more biliary plastic
or metal stents were introduced beyond the stricture to relieve biliary obstruction.
Clinical characteristics of the stricture were considered when making a decision on
the type of stent.
The cups of the biopsy forceps can be easily opened despite creating the bends. Rarely,
it may appear to open sluggishly but that may have more to do with the stricture than
the bends. However, the creations of the bends make it difficult to rotate. This cannot
be consistently achieved and therefore not routinely attempted. Rarely, flipping maybe
attempted to angulate the forceps into one of the intrahepatic ducts with limited
success.
Follow-up
Malignant cases were referred to oncological evaluation and/or surgery for further
treatment depending on the stage. If the biopsies were negative, FSEB was repeated
in 1 to 2 months if there was a clinical concern for malignancy. Rarely, if the suspicion
of malignancy was very high, patients were referred to be evaluated for surgery despite
negative endobiliary biopsies. If the suspicion for malignancy was low or moderate,
multiple endobiliary biopsies were performed at the time of stent change.
Final diagnosis
Biliary biopsy samples were routinely classified into one of the following categories:
(1) benign lesion, (2) adenoma, (3) dysplasia, (4) malignant lesion, (5) suspected
malignant lesion, or (6) insufficient material. The samples that were labeled as “insufficient
material” were then considered a technical failure. The biopsy samples interpreted
by the pathologist as “adenoma”, “dysplasia”, “malignant” or “suspected malignant”
were considered to be neoplastic by our definition. All biopsies with atypical pathologic
results were considered benign. A benign biopsy was corroborated with a compatible
clinical course and follow-up for at least 1-year.
The final diagnosis of these samples was made either on the basis of surgical pathology,
subsequent EUS-FNA or ERC sampling with definite evidence for malignancy or a benign
long-term radiological follow-up for 1-year. The neoplastic samples identified by
FSEB were considered true positives. Likewise, the benign samples with a benign clinical
course of 1-year with a stable or improved radiological imaging were deduced to be
true negatives. These patients were further reviewed until their most recent clinical
encounter including all interval radiographic imaging, procedures/surgeries, pathologic
data, and deaths (verified through obituaries and medical records). The benign samples
that were subsequently found to be malignant, either by surgical pathology, EUS-FNA,
repeated FSEB, or clinical evidence of metastasis in the areas of the stricture were
considered to be false negative. Patients who dropped out within one year of follow-up
were considered false negative by our strict definition. For the per protocol analysis,
the drop-out patients were excluded.
Data collection
Data was collected by reviewing electronic medical records, including the medical
history, details of the procedure, location of the stricture, complications, histopathological
reports, surgical procedure reports and follow-up information. The data were then
compiled using Microsoft Excel (Microsoft Corporation, Redmond, Washington, United
States).
Statistical evaluation
The primary endpoint was the diagnostic accuracy of FSEB based on per patient analysis.
Secondary outcomes were subgroup analysis of diagnostic accuracy of FSEB based on
the site of stricture and type of malignancy. Statistical analysis was performed using
SPSS software (IBM, Chicago, Illinois, United States). Continuous variables are expressed
as mean ± standard deviation (SD), or as median with range, if data are non-parametric.
The diagnostic characteristics were calculated as a proportion and presented as accuracy,
sensitivity, specificity, and negative predictive values (NPV). Fisher’s exact test
was used to compare the differences in demographics between patients with false-negative
FSEB and all other patients. Multivariate logistic regression analysis was performed
to evaluate for independent predictors of false-negative results. P < 0.05 was considered to be statistically significant.
Per-biopsy session analysis
A secondary analysis was performed to evaluate the diagnostic characteristics of FSEB
based on per biopsy session analysis. This analysis is performed such that the denominator
for the calculations was the total number of sessions of biopsy rather than the total
number patients. The goal of this analysis is to show the diagnostic characteristics
of FSEB per ERCP biopsies. Sensitivity, specificity, accuracy, and negative predictive
value were calculated. Comparative analysis of the yield of FSEB was made based on
the location of the biliary stricture.
Results
Demographics
A total of 227 patients undergoing endoscopic biliary biopsies for evaluation of biliary
stricture were initially enrolled in this study. Twenty-three patients met the exclusion
criteria and were excluded from the analysis (10 underwent choledochoscopic biopsy
only; 8 patients died of unrelated reasons or chose to terminate follow-up due to
advanced age or severe co-morbidities; 5 were sampled for other reasons) ([Fig. 1 ]). Final analysis was performed on 204 patients. Of these, six patients (2.9 %) received
post Roux-en-Y surgical procedures (4 for pancreaticobiliary cancers, 1 for autoimmune
cholangiopathy, 1 for bariatric surgery), eight patients (3.9 %) underwent choledochoscopy
in the same session. All of them underwent FSEB. The flow chart of patient enrollment
with inclusion and exclusion criteria is shown in [Fig. 1 ]. Neoplasia was identified in 117 patients (57.4 %), and 77 patients (37.7 %) had
no evidence of dysplasia over ≥ 1 years of follow-up ([Table 1 ]). Overall, 10 patients (4.9 %) were lost in follow-up.
Table 1
Demographic and final diagnoses of the patients included in the study (N = 204).
Characteristics
Value
Age in years, mean (range)
65 (31–100)
Sex, male (%)
113 (55.4 %)
Location of stricture
97
107
Biopsy session
166
38
Post Roux-en-Y surgery
6
Choledochoscopy or spyglass in the same session
8
Final diagnosis
77
19
13
9
9
6
21
Neoplasia
117
8
46
44
4
15
Dropout before 1-year follow-up
10
Diagnostic characteristics of FSEB
All tissue specimens obtained by FSEB were adequate for histopathological diagnosis.
Among these, a total of 103 patients were diagnosed with neoplasia (95 with cancer,
8 with adenoma/dysplasia) and were considered to be true-positive findings. Ninety-one
patients had non-neoplastic index FSEB histology: 77 (5 confirmed on surgery) had
no evidence of dysplasia upon follow-up of one year (true negative). Fourteen patients
were found to have neoplastic disease subsequent to their negative tissue sampling
and the index histology in these patients was considered to be false negative. Of
these, eight were found to have neoplasia on repeat EUS-FNA, five on surgical histopathology,
and one had an obvious source of malignancy with a cervical cancer history and multiple
liver metastasis on imaging. Ten patients dropped out before the 1-year follow-up
period and were categorized as false negative by our strict definition ([Table 1 ]).
The diagnostic characteristics of FSEB are summarized in [Table 2 ]. The overall sensitivity, specificity, accuracy, and NPV were 81.1 %, 100 %, 88.2 %,
and 76.2 %, respectively. When the dropout patients were excluded from the analysis
(per protocol), the sensitivity and accuracy improved to 88.0 % and 92.7 %, respectively.
Potential complications of FSEB include biliary duct perforation, bile peritonitis,
bile leaks, bleeding, infection, and pancreatitis. No significant complications were
noted on review of these patients.
Table 2
Diagnostic characteristics of FSEB for diagnosis of biliary neoplasia.
Total numbers
True positive
False positive
True negative
False negative[1 ]
Sensitivity (95 % ci)
Accuracy (95 % ci)
Negative predictive value (95 % CI)
Per-patient analysis (N = 204)
All locations
204
103
0
77
24
81.10 % (73.20 %–87.50 %)
88.24 % (83.00 %–92.31 %)
76.24 % (66.74 %–84.14 %)
Proximal
97
51
0
41
5
91.07 % (80.38 %–97.04 %)
94.85 %[2 ] (88.38 %–98.31 %)
89.13 % (76.43 %–96.38 %)
Distal
107
52
0
36
19
73.24 % (61.41 %–83.06 %)
82.24 %[2 ] (73.67 %–88.96 %)
65.45 % (51.42 %–77.76 %)
Per-biopsy session analysis (N = 250)
All locations
250
114
0
104
32
78.08 % (70.49 %–84.50 %)
87.20 % (82.41 %–91.08 %)
76.47 % (68.44 %–83.32 %)
Proximal
122
59
0
55
8
88.06 % (77.82 %–94.70 %)
93.44 %[2 ] (87.49 %–97.13 %)
87.30 % (76.50 %–94.35 %)
Distal
128
55
0
49
24
69.62 % (58.25 %–79.47 %)
81.25 %[2 ] (73.40 %–87.60 %)
67.12 % (55.13 %–77.67 %)
FSEB, fluoroscopy-guided, shaped endobiliary biopsy.
1 Including patients dropped out in follow-up (N = 10).
2 Accuracy rates between proximal and distal bile duct biopsies are statistically different
(P < 0.01).
Differences in diagnostic characteristics of FSEB based on malignancy and location
Because the diagnostic accuracy of endobiliary biopsies can be affected by the location
of the stricture [10 ], we compared the diagnostic performance characteristics of FSEB between proximal
and distal strictures. As shown in [Table 2 ], if the dropout subjects were included, FSEB detection for proximal biliary lesions
was significantly more sensitive (91.1 % vs 73.2 %, P < 0.01) and more accurate (94.9 % vs 82.2 %, P < 0.01) than for distal biliary lesions. When further dividing the stricture site
into intrahepatic duct (IHD), common hepatic duct (CHD), and common bile duct (CBD),
the sensitivity was 88.9 %, 91.5 %, and 73.2 % respectively ([Table 3 ]). FSEB had a significantly higher accuracy for lesions in the CHD compared to CBD
(94.5 % vs 82.2 %, P < 0.01).
Table 3
Diagnostic performance of FSEB according location or type of malignancy.
Patients (n)
Sensitivity (95 % CI)
NPV (95 % CI)
Accuracy (95 % CI)
Site of Stricture
Intrahepatic duct
24
88.89 % (51.75 %–99.72 %)
93.75 % (69.77 %–99.84 %)
95.83 % (78.88 %–99.89 %)
Common hepatic duct
73
91.49 %[1 ] (79.62 %–97.63 %)
86.67 % (69.28 %–96.24 %)
94.52 %[1 ] (86.56 %–98.49 %)
Common bile duct
107
73.24 %[1 ] (61.41 %–83.06 %)
65.45 % (47.26 %–80.90 %)
82.24 %[1 ] (73.67 %–88.96 %)
Malignancy
Cholangiocarcinoma
46
91.3 % (79.21 %–97.58 %)
95.06 % (87.84 %–98.64 %)
96.75 % (91.88–99.11 %)
Pancreatic Cancer
44
84.09 % (69.93–93.36 %)
91.67 % (83.58 %–96.58 %)
94.21 % (88.44 %–97.64 %)
FSEB, fluoroscopy-guided, shaped endobiliary biopsy.
1 Sensitivity and accuracy rates between common hepatic duct strictures and common
bile duct strictures are statistically different (P < 0.01 and P < 0.02, respectively).
The diagnostic performance of FSEB was also evaluated based upon type of malignancy.
The sensitivity for detecting cholangiocarcinoma was 91.3 % in comparison to pancreatic
cancer, which was 84.1 % ([Table 4 ]). Moreover, the accuracy was high in both groups, with cholangiocarcinoma having
a 96.8 % accuracy and pancreatic cancer with 94.2 % accuracy. As expected, the specificity
and positive predictive value was 100 % in both groups.
Table 4
Characteristics of confirmed false-negative results of FSEB (N = 14).
Multivariate analysis results
Characteristics
Value
Standardized β
% explained variance
P value
Age in year, mean (range)
60 (40–78)
0.0026
0.4
0.8263
Gender, male (%)
11 (78.6 %)
0.0843
1.2
0.2573
Location of stricture
5/51
0.0435
1.3
0.7601
9/52
FSEB, fluoroscopy-guided, shaped endobiliary biopsy.
Per-biopsy session analysis
Next, we analyzed the diagnostic characteristics of FSEB per biopsy session. A total
of 250 biopsy sessions were performed in the group of 204 patients. The overall sensitivity,
specificity, accuracy, and NPV were 78.1 %, 100 %, 87.2 %, and 76.5 %, respectively
([Table 2 ]). Similarly, compared to distal biliary lesions, FSEB was significantly more sensitive
(88.1 % vs 69.6 %, P < 0.01) and more accurate (93.4 % vs 81.3 %, P < 0.01) in diagnosing proximal neoplastic biliary lesions, by the strict definition
([Table 2 ]).
Independent predictors of False-negative FSEB
The demographics, clinical indication, location of FSEB, and final diagnosis for the
patients who were false negative on index FSEB are presented in [Table 4 ]. False-negative results were identified in 11.1 % (14/117) of the patients with
confirmed neoplasia. Of the 14 false-negative cases, four primary biliary malignances
(cholangiocarcinoma) and 10 secondary biliary malignances were subsequently diagnosed.
The secondary biliary malignancies included seven pancreatic cancers, two lymphomas,
and one metastatic cervical cancer. According to multivariate analysis, none of the
variables we tested, such as age, gender, location of lesion, or tumor type, were
significant independent predictors of false negative by FSEB. It should be noted that,
under the strict definition where the ten patients dropped out in the follow-up were
included for analysis, the false-negative rate is higher for lesions in the distal
bile duct than the ones in the proximal region (Fisher’s exact test; P = 0.01).
We also reviewed the records of the 77 negative patients beyond the 1-year follow-up
period to identify potential misdiagnosis. Eleven patients (11/77; 14.3 %) did not
return to our institution for care beyond one year, one of whom was found to have
died of unknown cause on extensive review of death and obituary records. The remaining
66 patients (66/77; 85.7 %) were reviewed with a mean follow-up of 89.7 months (range,
24–223 months). At 2 years following FSEB, all 66 patients continued to have a benign
course and none were reclassified to false negative. One patient developed periampullary
adenocarcinoma after 30 months, which could possibly be interpreted as a false negative.
Death records and obituary data were reviewed on all 66 patients and 10 individuals
had passed after a mean of 69.3 months (range, 25–154) from date of their procedure.
Discussion
In this long-term, retrospective study of a large cohort of patients, we show very
high overall sensitivity and accuracy of FSEB for the diagnosis of biliary neoplasm
(81.1 % and 88.2 %, respectively). In comparison, the sensitivity of conventional
fluoroscopic forceps biopsies ranges from approximately 36 % to 81 % with the pooled
sensitivity of 48 % [11 ]. The sensitivity of biliary brush cytology varies from 6 %–64 % with the overall
sensitivity of 42 % [12 ]. The combination of brush cytology and forceps biopsy has been shown to enhance
the yield of biliary malignancy to about 70 % to 74 % [13 ]
[14 ]. Repeated brush cytology and biopsies can moderately increase the sensitivity [15 ]
[16 ]. An ERC-based method of obtaining pathologic diagnosis using a smashed cytologic
preparation of forceps biopsy sampling (Smash protocol) showed an overall sensitivity
of 76 % [17 ]. Thus, our results show that the sensitivity of FSEB is among the highest of what
has been reported.
Our data show that the sensitivity and accuracy of FSEB for detecting proximal biliary
neoplasm were significantly higher than that for distal biliary neoplasm, which is
consistent with a previous study [18 ]. The disparity is conceivably due to the intrinsic difficulty to access and target
a lesion located proximally in bile ducts.
Modifying the design of biopsy tools represents an important strategy to improve diagnostic
yield. The development of novel biopsy forceps or biopsy methods, such as the use
of double balloon enteroscopy forceps, the “ropeway-type forceps” with a side slit
for a guidewire, and the 90° adjustable biopsy forceps, has been reported and shown
to improve the sensitivity up to 69 % to 71 % [6 ]
[7 ]
[8 ]. However, these studies only involve small groups of subjects (12–43 patients) [6 ]
[7 ]
[8 ]. Thus, up until this study, high quality studies that establish the usefulness of
a specific biopsy forceps or technique have been limited. To the best of our knowledge,
our study has included one the largest cohorts for evaluation of the diagnostic characteristics
of ERC-based biopsy in biliary neoplasms.
A unique feature of FSEB is the use of biopsy forceps that have been precurved at
the tip and bent further back to mimic the shape of the bile duct in question ([Fig. 2 ], [Fig. 3 ] and [Video 1 ]). The “shaped” forceps provide four distinct advantages. First, the angulated tip
allows smooth exit from the duodenoscope elevator. Second, it improves the ease of
passage through the papilla. Third, the shaped forceps enhance targeting of the lesion,
regardless of the shape and distance above the papilla. Fourth, by altering the degree
of tightness of forceps closure, the instrument can negotiate through a tortuous bile
duct to reach the stricture. It should also be noted that our approach features the
standard forceps with a cup opening size of 7.1 mm (the model, producer had been shown
in methods), which have been shown to be superior to the pediatric or the SpyBite
forceps that have smaller cup opening sizes (4.9 mm and 4.1 mm, respectively) [19 ]
[20 ]. The ability to exert moderate pressure during closure is an integral part of the
FSEB technique, which can only be carried out with a strong, standard forceps. For
this reason, forceps with a thinner or softer shaft are less optimal in acquiring
the desired tissue. With the modification, we have been able to insert the biopsy
forceps to the papilla in all patients intended (i. e. post-sphincterotomy patients).
Overall, we believe that the ability to create bends in the forceps and the use of
the tension to help with angulation allows for easier access of the bile duct and
better targeting of the tumor, which contributes to higher sensitivity.
The routine usage of a guidewire across the stricture to guide forceps advancement
and tissue targeting is another essential element of our biopsy technique, as fluoroscopic
guidance based on the contrast retained above the stricture is unreliable with regards
to safety and accuracy of forceps passage. The shaped biopsy forceps and the guidewire
can be simultaneously inserted in one channel. No false-positives or complications
have been found with FSEB. Based on our experience, we believe that FSEB may have
a lower risk of complications compared to standard biopsy forceps as the shaped forceps
better align with the bile duct. We did not conduct any analysis on the safety and
accuracy of FSEB before or after balloon dilation of the stricture, although anecdotally
we have not observed any difference between the two approaches.
The high diagnostic yield of FSEB can, in part, be attributed to the increased amount
of tissue sampled from the stricture. A previous study showed that the tissue sampling
number is associated with lower false-negative rate and recommended at least 4–6 samples
per case [21 ]. In the FSEB study, we generally obtain the equivalence of three adequately sampled
biopsies, with an additional sampling performed if one of the samples was substantially
smaller than expected. The pathological tissue samples collected by FSEB range from
1 mm to 7 mm and were sufficient for histopathological diagnosis in all the patients.
We were not able to identify any statistically significant independent predictors
of false-negative results by FSEB. The sensitivity for all sampling methods can be
influenced by the characteristics of the tumor underlying the stricture [10 ]
[22 ]
[23 ]. In our study, the false negative rate for the secondary malignant biliary stricture
was moderately higher than that of cholangiocarcinoma.
The ropeway technique from multiple vendors, as well as several other kinds of “modified”
biliary forceps, has been attempted by our senior author over the years. Despite the
senior author’s experience in ERCP for more than four decades, it is unclear why our
biopsy technique works better than the others. It is possible that the cups of the
biopsy forceps have better ability to obtain adequate tissue. Other factors, such
as the stiffness of the forceps, the size of forceps, and the smoothness of the edges,
may also play a role.
We recognize several limitations of the study. First, this was a retrospective, single-centered
study and future prospective studies are needed to confirm the advantage of FSEB.
Second, we have compared the accuracy of FSEB with that of previously reported ERC-based
methods ([Table 5 ]). The results remain to be confirmed by further studies that include a control group
using a conventional biopsy forceps. Third, operator-dependent bias may exist, as
all FSEB were performed by a single endoscopist. There may be a learning curve with
this method, but performance may improve with more procedures. This study did not
evaluate the role of skill or learning curve on the accuracy of biopsy. It would be
of importance to determine whether the results can be reproduced by other endoscopist
in future studies. Fourth, although we employed a scheme of 1-year follow-up to determine
the final diagnosis of non-neoplastic cases, we recognize that biliary adenoma without
dysplasia or with low grade dysplasia can be slow-growing and may evade definite diagnosis
in the one-year period, leading to an underestimate of the false-negative rate. For
this reason, we continued to review the patients beyond the standard 1-year follow
up corroborating death records to obviate the need for potential reclassification.
Table 5
Characteristics of studies evaluating the sensitivity of endobiliary forceps biopsies
and techniques.
Study
Year
Sample size (n)
Technique
Sensitivity %
Sugiyama et al. [24 ]
1996
52
Forceps biopsy
81 %
Kitajima et al. [25 ]
2007
51
Forceps biopsy
65 %
Weber et al. [26 ]
2008
58
Forceps biopsy
53 %
Writght et al. [17 ]
2011
133
Endobiliary forceps biopsy + Smash Protocol
72 %
Hartman et al. [27 ]
2012
81
Forceps biopsy
76 %
Draganov et al. [28 ]
2012
26
Endobiliary forceps biopsy
29 %
Weilert et al. [29 ]
2014
51
Brush cytology + forceps biopsy
50 %
Chen et al. [30 ]
2016
79
Endobiliary forceps biopsy
54 %
Yamamoto et al. [31 ]
2017
360
Endobiliary slim forceps biopsy
70 %
Inoue et al. [32 ]
2018
110
Controllable biopsy-forceps
60 %
Our study
2021
204
Endobiliary shaped forceps biopsy
81 %
Conclusions
In summary, our results indicate that FSEB exhibits high levels of sensitivity and
accuracy in evaluating biliary strictures and is safe and inexpensive. FSEB can potentially
be used solely as the tissue sampling method. Future, prospective, randomized, multicenter
studies are needed to validate the definitive role of FSEB in accurate diagnosis of
malignant biliary neoplasm. Further studies are also needed to compare FSEB with other
sampling techniques, such as biliary brush cytology, cholangioscopy, and EUS-FNA/biopsy,
in order to establish a highly sensitive and specific diagnostic modality for biliary
stricture that is also safe and cost-effective.
