Introduction
Endoscopic forceps biopsy has long been the standard in luminal endoscopy, with high
accuracy for epithelial lesions, but tissue diagnosis during endoscopic retrograde
cholangiopancreatography (ERCP) remains less accurate and less well standardized [1 ]
[2 ]. Advanced imaging techniques such as confocal microscopy [3 ] and cholangioscopic visual assessment using artificial intelligence [4 ], as well as endoscopic ultrasound (EUS)-guided sampling [5 ], have been added to the spectrum. Multiple methods are often combined to reach a
definitive diagnosis [6 ].
Diagnostic accuracy for biliary strictures remains variable, particularly for detecting
or ruling out malignancy, which carries major clinical implications. Most available
data come from retrospective case series – some comparative, few prospective or randomized
– with variable results across techniques [7 ]
[8 ]. A 2015 meta-analysis [9 ] and a 2023 ASGE survey [1 ] reported sensitivities of 40%–60% for intraductal biopsy and brushing, while cholangioscopic
biopsies showed slightly better performance (70%–75%). Given the demands of personalized
cancer therapy and the need for detailed histologic analysis, a more reliable intraductal
biopsy method is needed, ideally matching the effectiveness of standard luminal biopsies.
Cryobiopsy is well established for tissue sampling in bronchoscopy [10 ]
[11 ]
[12 ], and a similar probe has shown promising results in the biliary system in ex vivo
studies [13 ] and a single patient case [14 ]. We now report the first multicenter pilot series comparing cryobiopsy with percutaneous
cholangioscopic forceps biopsy. The cryoprobe, similar to its bronchoscopic counterpart,
is evaluated here as a proof of principle before its potential development for ERCP
use. Because of its limited length, only percutaneous application was feasible.
Methods
Patients and procedures
Patients undergoing percutaneous cholangioscopy with tissue sampling for biliary strictures
were enrolled at five centers between March (first treatment) and August 2024 (last
visit). Institutional Review Board approval was obtained at all sites. Patients were
included if they had a biliary stricture without a mass lesion or clear diagnosis
based on their history or a prior biopsy, and a tissue diagnosis was still required
for further management. This also included patients with primary sclerosing cholangitis
(PSC) presenting with a dominant stricture, where access via ERCP or enteroscopy was
not feasible. Percutaneous drainage was established prior to cholangioscopy as necessary.
Indications, plus patient and procedural data, including cryo activation time, number
of biopsy attempts, gas pressure, procedure duration, and biopsy area, were recorded.
Cholangioscopy was performed using a single-use cholangioscope (SpyGlass Discover;
Boston Scientific, USA) that was introduced via the percutaneous tract according to
each center’s standard procedure, typically using a sheath and/or a guidewire. The
cholangioscopic visual appearance was used to guide biopsies. The biopsy techniques
were applied in randomized order (sealed envelopes), with either six forceps biopsies
or three cryobiopsies performed first. Each sample was placed in a separate pathology
vial (nine per patient). A post-biopsy drain was routinely inserted, as per the standard
practice at all participating centers.
Forceps biopsy
SpyBite Max biopsy forceps (Boston Scientific) were used for all forceps biopsies.
Any additional biopsies taken beyond the six study samples were excluded from the
analysis and were used solely for clinical purposes.
Cryobiopsy
The cryobiopsy technique is similar to that used in bronchoscopy [10 ]; the equipment used is described in Table 1s , see online-only Supplementary material. A 1.1-mm flexible cryoprobe ([Fig. 1 ]) was connected to the ERBECRYO 2 unit, which automatically recognized the probe
and selected the appropriate settings. Functionality was tested in sterile water with
a 5-second activation via foot pedal. For biopsy, the probe was positioned tangentially
or perpendicularly to the target lesion and activated for 4–5 seconds, guided by an
acoustic signal. It was then withdrawn, together with the cholangioscope, through
the percutaneous tract, remaining activated until the specimen was released into fluid
and then formalin. This process was repeated twice, for a total of three attempts,
even if no tissue was obtained ([Video 1 ]).
Fig. 1 The ERBECRYO 2 device with its corresponding single-use 1.1-mm cryoprobe.
The cryo principle is demonstrated using a water container, where the frozen attachment
is visible. Percutaneous cholangioscopy is then performed in a patient with intrabiliary
projections to diagnose or rule out cancer. The probe is introduced and pressed against
the lesion; after the foot pedal has been pressed for 4–5 seconds, the probe and cholangioscope
are removed together, with a large attached specimen clearly visible.Video 1
Histopathology
Biopsies were evaluated by a specialized gastrointestinal (GI) histopathologist at
each center, who was blinded to the acquisition method. For intraindividual comparison,
the same standard histopathologic criteria were applied to both biopsy types from
each patient.
Follow-up
Follow-up was conducted before discharge and on day 30 ± 7 via telephone. Patients
were asked about any symptoms potentially related to the procedure (e.g. bleeding
or pain), as well as any medical consultations or hospital admissions. To ensure the
diagnosis of benign biliary stricture, follow-up was subsequently conducted clinically
and by imaging for at least 6 months.
Outcomes and definitions
Main objective
The main objective was the feasibility of percutaneous cryobiopsy in the bile duct,
meaning the successful retrieval of at least one cryobiopsy specimen per patient.
Further objectives
Success rate Percentage of successful biopsy retrievals per technique, based on total attempts
(45 cryobiopsies and 90 forceps biopsies).
Biopsy volume Estimated by the overall mean area (mm2 ) per biopsy, calculated by summing all section areas per biopsy and patient (2D surrogate
for 3D volume).
Biopsy quality Assessed using a histopathologic Likert scale for each technique (adapted from Wirsing
et al. [13 ]).
Representativeness Defined as a score >2, indicating moderate limitations but sufficient quality to
assess relevant morphologic and histologic features.
Crush artifacts Graded by the percentage of artifact area per sample: none; low, 1%–19%; moderate,
20%–39%; or high, >40%.
Procedure time Measured overall and separately for each biopsy method.
Adverse events Adverse events included death, bleeding, post-procedural bleeding, perforation, infection,
and abscess formation. Events were categorized according to ISO 14155:2020 and MDR
2017/745 (Articles 2 and 80), and graded by severity (mild, moderate, severe).
Statistics
For descriptive statistics, categorical variables are presented as counts and percentages,
and continuous variables as mean (SD) or median with interquartile range (IQR; 25th–75th
percentile). Normally and non-normally distributed data were compared using Student’s
t test and Mann–Whitney U test, respectively. Differences in means or medians are reported with 95%CIs. A two-sided
P value <0.05 was considered statistically significant. Analyses were performed using
GraphPad Prism, version 9.5.1 (GraphPad Software LLC).
As this was the first in vivo feasibility study, no formal power or sample size calculation
was performed. A sample of 15 patients was chosen to assess the feasibility of percutaneous
bile duct biopsy and inform future large randomized multicenter trials.
Safety assessment and reporting requirements
Participant safety was monitored throughout the study. The investigation was approved
by the Ethics Committee of the University of Tübingen (No. 401/2023MP1), was performed
according to the Helsinki declaration, and was reported to the German Federal Institute
for Drugs and Medical Devices (BfArM; EUDAMED No. CIV-23–06–043299), in accordance
with the Medical Device Law Implementation Act (MPDG) and EU MDR 2017/745.
Results
A total of 15 patients were included (7 men; mean age 60.2 [SD 18.5]; body mass index
23.1 kg/m2 ) ([Table 1 ]). The five centers enrolled seven, four, two, one, and one patients each. The mean
(SD) total procedure time was 13.58 (7.28) minutes, defined from cholangioscope insertion
to the final biopsy. The mean (SD) times were 7.14 (2.40) minutes for forceps biopsy
and 6.45 (6.45) minutes for cryobiopsy (P = 0.74).
Table 1 Demographic and procedural data for the 15 patients with undiagnosed biliary stricture
who underwent biopsies using forceps and cryobiopsy.
All
Forceps biopsy
Cryobiopsy
95%CI for mean difference
P value
1 Six biopsies were taken by forceps (total of 90) and three by cryobiopsy (total of
45).
2 For details of the determination of the specimen being representative (yes/no) and
for area calculation as a surrogate parameter for volume see Methods.
3 Graded 0 (none) to 3 (high).
4 Likert scale, 0–6.
Age, years
60.2 (18.5)
Sex, female, n
8
Procedural time, mean (SD), minutes1
13.6 (7.3)
7.1 (2.4)
6.5 (4.6)
−2:3 to 3:3
0.74
Histologic results
Adequate samples, %
100 (15/15)
100 (15/15)
>0.99
96 (87/90)
93 (42/45)
0.40
Representative specimens obtained, %2
74.70 (65/87)
97.60 (41/42)
0.001
Area, mean (SD), mm2 2
1.87 (1.86)
8.54 (6.00)
−8.07 to −5.27
<0.001
Crush artifacts, median (IQR)3
1 (1)
1 (1)
0 to 1
0.06
Artifact-free areas, mean (SD), %
85.5 (19.3)
93.5 (7.0)
−14.1 to −1.9
0.01
General assessment score, median (IQR)4
4 (3)
5 (2)
−2 to 0
<0.001
Histology results and final diagnoses are summarized in [Table 1 ] and [Fig. 2 ]. At least one adequate sample was obtained in all patients and by both biopsy techniques.
Adequate specimens were retrieved in 42/45 (93.3%) cryobiopsies and 87/90 (96.7%)
forceps biopsies. Cryobiopsies yielded significantly larger (8.54 [SD 6.00] vs. 1.87
[SD 1.86] mm2 ; P < 0.001) and more representative samples (97.6% vs. 74.7%; P = 0.001), with a higher percentage of artifact-free areas (93.5% [SD 7.0%] vs. 85.5%
[SD 19.3%]; P = 0.01) and higher histology scores on the 6-point Likert scale (5 [IQR 2] vs. 4
[IQR 3]; P < 0.001). Notably, the first cryobiopsy was representative in 100% of cases vs. 57.1%
for the first forceps biopsy. Representative images are shown in [Fig. 3 ] and [Fig. 4 ]; clinical and histologic data are detailed in [Table 2 ]. There was no significant difference in biopsy size based on the sequence of the
biopsy technique. Neither cryobiopsy (first, 7.5 [SD 4.7] mm2 ; second, 9.3 [SD 9.3] mm2 ; P = 0.34) nor forceps biopsy (first, 1.6 [SD1.8] mm2 ; second, 2.2 [SD 2.0] mm2 ; P = 0.17) showed statistically significant differences (Fig. 1s ).
Fig. 2 Graphical comparison of cryobiopsy and standard forceps biopsy with cryobiopsy resulting
in: a significantly larger biopsies; b more representative specimens; c specimens with a higher histopathologic assessment score.
Fig. 3 Example macroscopic specimens of biliary tissue obtained using: a the cryoprobe; b forceps biopsy via percutaneous cholangioscopy.
Fig. 4 Histologic appearance of the obtained tissue: a,b displayed on the slides; c,d at 1.5× magnification; a,c forceps biopsy; b,d biopsy using the cryoprobe.
Table 2 Individual patient data.
Age, years; sex
Diagnostic setting
Final diagnosis
Diagnosis by:
Follow-up, months
Outcome
Forceps
Cryo
BDA, biliodigestive anastomosis; CBD, common bile duct; CCA, cholangiocarcinoma; CRC,
colorectal carcinoma; ERCP, endoscopic retrograde cholangiopancreatography; HCC, hepatocellular
carcinoma; IPMN, intraductal papillary mucinous neoplasm; LGIN, low grade intraepithelial
neoplasia; LHD, left hepatic duct; LTX, liver transplant; PSC, primary sclerosing
cholangitis; PTBDE, percutaneous transhepatic biliary drainage exchange.
1 Biopsy was indicated to exclude neoplasia in late strictures after liver transplantation
(LTX) for primarily benign disease under immunosuppression. 2 Imaging and biopsy negative, further follow-up pending.
81; male
BDA stenosis after pancreatectomy (high grade IPMN)
Inflammatory stenosis
X
X
3
Stricture resolution
59; female
Late BDA stenosis after LTX1
Inflammatory stenosis
X
X
12
Switched to ERCP
28; female
Late LHD stenosis after LTX (cryptogenic liver failure)1
Inflammatory stenosis
X
X
11
Stricture resolution
61; female
BDA stenosis after extended right hemihepatectomy (CCA)
Recurrence of CCA
X
X
4
Cancer death
73; female
BDA stenosis after right hemihepatectomy (PSC)
Inflammatory stenosis
X
X
3
Death from liver cirrhosis
19; female
Late stenosis of left HD after LTX1
Inflammatory stenosis
X
X
8
Switched to ERCP
59; male
BDA stenosis after left hemihepatectomy (LGIN)
Inflammatory stenosis
X
X
6
Stricture resolution
44; male
Late BDA stenosis in PSC
Inflammatory stenosis
X
X
11
Repeated PTBDE
77; male
Late biliary stenosis after LTX for HCC
Exclusion of malignancy
X
X
11
Repeated ERCP
76; female
BDA stenosis after pancreatectomy for cancer
Chronic ulcer2
X
X
8
Repeated PTBDE
67; female
BDA stenosis after surgery for CCA
Local recurrence
0
0
3
PTBDE, fatal ulcer bleeding
64; male
LHD stenosis after right hepatectomy for CRC metastases
Local recurrence
0
X
8
Palliative chemotherapy
77; male
Suspicious hilar stricture
CCA
X
X
3
Cancer death
45; female
Suspicious distal CBD stricture
Pancreatic carcinoma
0
0
–
Pancreatectomy, T1N1
73; male
BDA stricture after surgery for lymphoma
Inflammatory stenosis
X
X
7
Lymphoma, stable
All patients were assessed for adverse events at discharge and at 30 days. Three reported
post-procedural abdominal pain, two had fever, one developed cholangitis, and one
had a wound infection at the access site. None of the events were directly linked
to either of the biopsy techniques. No bleeding or perforation occurred. All adverse
events resolved with analgesics and/or antibiotics.
Discussion
Tissue diagnosis can be difficult in the biliary tract; even forceps biopsy under
direct vision with cholangioscopy may often not provide sufficient samples, in part
due to tissue hardness and the angle of biopsy. These factors help explain the well-known
limitations in obtaining sufficient tissue samples from the bile duct, where specimens
are often much smaller than the actual cup size of the forceps. This limitation was
one of the driving forces behind the development of cryobiopsy, which relies on tissue
freezing, adherence, and traction to retrieve samples. We present the results of a
pilot, proof-of-principle study on the use of biliary cryobiopsy for biliary strictures.
The findings are promising in terms of the technical feasibility and, notably, the
superior quality of the histologic specimens. Cryobiopsy yielded significantly larger
and more representative samples than standard forceps biopsy, with fewer artifacts
and higher histopathologic assessment scores.
To align with current clinical practice, twice as many forceps biopsies were performed;
however, cryobiopsy may ultimately achieve high diagnostic yields with fewer samples.
This could prove valuable for more detailed histologic analysis or when limited sampling
is preferred, particularly once ERCP-compatible cryoprobes are developed. For this
study, we used a short cryoprobe adapted from bronchoscopy to assess the initial feasibility
of the biliary cryobiopsy technique in general, and employed a randomized design to
ensure comparability with forceps biopsy.
Because of the limited sample size, our study could not assess diagnostic accuracy
or provide a comprehensive safety analysis. As both biopsy methods, forceps and cryoprobe,
were used in the same patients, attribution of any complications to a specific technique
is currently limited. No severe procedure-related complications (e.g. bleeding or
perforation) were observed. Safety will need further evaluation, particularly in the
ERCP setting.
Following this feasibility study on tissue yield, it is clear that further prospective
and comparative studies with adequate outcome parameters are needed to define the
clinical role of cryobiopsy in biliary strictures. Such future cryobiopsy studies
should consider pretest probability based on clinical context, such as patient age
or underlying conditions like PSC or secondary sclerosing cholangitis, as diagnostic
needs differ across these groups. This has often been overlooked, contributing to
variable accuracy and limited data on clinical impact.
Historically, the sensitivity of brush cytology and forceps biopsy under fluoroscopy
has ranged between 40% and 60%, as reported in reviews from 2015 [9 ] and 2023 [1 ]. With the wider use of cholangioscopy and direct biopsy, diagnostic yield has improved.
The 2023 ASGE survey reported a 27% incremental gain across four comparative studies
[1 ]. The role of artificial intelligence in enhancing cholangioscopic imaging, alone
or combined with clinical and laboratory data, remains an area for future exploration
[15 ]
[16 ]. There may be a correlation between tissue volume and diagnostic yield or accuracy,
but limitations are especially prominent in biliary malignancies with a high proportion
of inflammatory and fibrous components. Our results may be encouraging in that respect,
and further studies are warranted.
Currently, biliary tissue acquisition often involves multiple diagnostic methods to
achieve adequate clinical accuracy. With the rise of personalized medicine and advances
in chemo- and immunotherapies in both palliative and neoadjuvant settings, the demand
for detailed histologic analyses, including immunohistochemistry and tumor sequencing,
has grown [17 ]
[18 ]. In many cases, a simple malignancy diagnosis is no longer sufficient; more tissue
is needed for precise treatment planning. Biliary cryobiopsy, especially with future
ERCP-compatible probes, may offer an advantage by providing larger, higher quality
samples to meet these evolving clinical needs. This should be further investigated
using a longer probe compatible with ERCP in larger prospective multicenter trials
to generate meaningful data on clinical utility.
