Keywords Pancreatobiliary (ERCP/PTCD) - ERC topics - Strictures
Introduction
To improve stent patency and survival in patients with malignant biliary obstruction,
endobiliary radiofrequency ablation (RFA) is an effective local palliative treatment
for tumors within the bile duct [1 ]. It improves bile duct patency by directly delivering ablative thermal energy to
the intraductal tumor burden, thus inducing coagulation necrosis [2 ]
[3 ]. However, there is a potential risk of serious adverse events (AEs) such as bleeding
or perforation due to the uncontrolled excessive necrotizing effect of endobiliary
RFA [4 ]
[5 ].
A recently introduced temperature-controlled RFA system could increase the safety
and efficacy of RFA [2 ]. The bipolar electrodes of this system have a temperature sensor, which continuously
monitors the temperature of the ablation site to prevent tissue overheating and reduce
the risk of unintentional thermal injuries [2 ]
[6 ]
[7 ]. The system automatically switches off when the electrode reaches a preset target
temperature. The catheter is available in lengths of 11, 18, 22, and 33 mm, allowing
for a variety of procedures for strictures of different shapes and lengths.
Although endobiliary RFA can prolong stent patency, its efficacy in inoperable hilar
cholangiocarcinoma (CCA) is unclear [8 ]
[9 ]. Because the hilar bile duct has complex duct-vascular contacts and a thin wall,
RFA-induced fatal AEs are more likely than in the distal bile duct [8 ]
[9 ]
[10 ]. We aimed to evaluate the efficacy and safety of temperature-controlled endobiliary
RFA in patients with inoperable hilar CCA.
Patents and methods
Study design and patients
We reviewed an endoscopic database from one tertiary referral center to retrieve information
on consecutive patients with inoperable CCA between August 2020 and November 2022.
The inclusion criteria were age >19 years; histologically confirmed hilar CCA; ineligibility
for surgery because of advanced tumor stage or operative risks; Eastern Cooperative
Oncology Group (ECOG) performance status ≤3; and no previous treatment. The exclusion
criteria were technical failure of a transpapillary approach due to gastric-outlet
obstruction or surgically altered anatomy; presence of other malignancy; severe renal
or hepatic dysfunction (creatinine clearance rate <10 mL/min or prothrombin time activity
≤40%); bleeding tendency (platelet count <50,000 mm3 or international normalized ratio >1.5); and contraindications for endoscopic retrograde
cholangiopancreatography (ERCP).
All included patients with definite stricture on cholangiogram and intraductal ultrasound
(IDUS) were offered the opportunity to receive endobiliary RFA. Because endobiliary
RFA may be associated with longer procedure time and additional medical costs, some
patients did not receive endobiliary RFA and were studied as part of the stent-only
group. This study was approved by our Institutional Review Board (SCHBC 2020–05–015–007)
and was conducted in accordance with the Declaration of Helsinki. The participants
provided written informed consent before undergoing procedures. This trial was registered
at https://cris.nih.go.kr (KCT0008576).
Temperature-controlled RFA system
The ELRA (STARmed, Goyang, South Korea) 7F RF catheter of 175 cm length has a bipolar
electrode (11, 18, 22, and 33 mm in length) with a temperature sensor at the distal
end, and can be connected to an RF generator (VIVA combo; STARmed) to deliver energy
in temperature-control mode ([Fig. 1 ]). The RF generator prevents excessive heating at the ablation site by automatically
switching heating on and off based on the predetermined target temperature. The RF
catheter can be inserted into the bile duct over a 0.025- or 0.035-inch guidewire
and generates thermal energy at an appropriate temperature to induce coagulation necrosis
of the obstructive lesion.
Fig. 1 Endobiliary radiofrequency (RF) catheter (ELRA; STARmed, Goyang, South Korea) and
power generator (VIVA Combo; STARmed). a Four types of RF catheter with ablation lengths of 11, 18, 22, and 33 mm. b The RF power generator and its settings. Temperature was continuously monitored during
the procedure.
Procedures
All procedures (including endobiliary RFA and stent placement) were performed by two
experienced endoscopists (J.H.M. and Y.N.L.). After being administered prophylactic
antibiotics, patients were placed in prone position with electrocardiogram monitoring
and supplied with oxygen. During ERCP procedures using a side-viewing duodenoscope
(TJF-260V; Olympus Medical Systems, Tokyo, Japan), carbon dioxide (Colosense CO-3000;
Mirae Medics Co., Seoul, South Korea) was used for insufflation to prevent AEs. After
successful selective bile duct cannulation and bilateral guidewire placement, a cholangiogram
was obtained under fluoroscopic guidance to identify the location and length of the
hilar stricture. IDUS using an ultrasonic probe (20 MHz/6F, transducer, UM G20 ± 29R;
Olympus Medical Systems) was performed to assess the stricture and the maximum thickness
of the corresponding bile duct wall. Balloon sweeping was conducted to confirm the
length of the stricture and remove any biliary sludge or stones that might interfere
with the RFA procedure.
In patients assigned to the RFA + stent group, the radiopaque electrodes of the RF
catheter preselected according to the characteristics of the stricture were advanced
over the guidewire and positioned at the stricture under fluoroscopic guidance. Temperature-controlled
endobiliary RFA was performed at an RF power of 7 W, a target temperature of 80°C,
and for a duration of 90 to 120 seconds [6 ]
[11 ]. Fluoroscopic views were obtained intermittently during the RFA procedure to ensure
that the entire target area received RFA. Two or more fractionated RFA applications
were attempted in the proximal-to-distal direction in an effort to ablate all lesions.
For instance, one application of RFA was sufficient in Bismuth–Corlette type I patients,
whereas two could be required to ablate the common hepatic duct (CHD) and one of the
left or right hepatic duct in Bismuth-Corlette type II-III patients. Similarly, some
Bismuth-Corlette type III-IV patients required three RFA applications in the CHD and
both the left and right hepatic ducts ([Fig. 2 ]). The duration of RFA was 120 seconds; during RFA for the right hepatic duct, this
was adjusted to 90 seconds based on the proximity of the right hepatic duct and right
hepatic artery. After RFA procedures, the RF catheter was removed from the bile duct
and self-expandable metal stents (SEMSs) ([Fig. 3 ]) or plastic stents (PSs) ([Fig. 4 ]; [Video 1 ]) were deployed to cover the stricture for palliative biliary drainage. Bilateral
stenting was performed for drainage in patients with Bismuth III-IV hilar CCA.
Fig. 2 Schema of radiofrequency ablation (RFA) application for hilar cholangiocarcinoma according
to the Bismuth classification.
Fig. 3
a Cholangiogram and b intraductal ultrasound to assess stricture. c Endoscopic view showing a radiofrequency ablation (RFA) catheter entering the bile
duct. d,e Fluoroscopic view showing bilateral endobiliary RFA using 18-mm RF catheters for
120 seconds at a power of 7 W and temperature of 80°C. f Successful deployment of metal stents after RFA.
Fig. 4
a Cholangiogram and b intraductal
ultrasound to assess stricture. c Endoscopic view showing a
radiofrequency ablation (RFA) catheter entering the bile duct. Fluoroscopic view showing
endobiliary RFA procedures using 11 mm RF catheters in the d
right intrahepatic duct (IHD) for 90 seconds, e left IHD for
120 seconds, and f common hepatic duct for 120 seconds at a
power of 7 W and temperature of 80°C. g Successful deployment
of a plastic stent after RFA.
Fractionated endobiliary radiofrequency ablation at a power of 7 W and temperature
of 80°C with subsequent deployment of plastic stents.Video 1
In patients assigned to the stent-only group, procedures including balloon sweeping
and stent placement, but not RFA, were performed as in the RFA + stent group. In both
groups, the decision about which or how many stents to place was at the discretion
of the endoscopist based on patient life expectancy and stricture severity. We used
uncovered braided nitinol SEMSs with a cross-wired structure (Bonastent M-hilar; Standard
Sci-Tech Inc., Seoul, South Korea), allowing for stent-in-stent deployment, and a
7F PS in a pigtailed configuration (Zimmon; Cook Medical., Winston-Salem, North Carolina,
United States) [12 ].
Data collection and follow-up
All patients underwent blood tests and simple abdominal X-rays at 4 hours, 24 hours,
and 1 week post-procedure, together with confirmation of clinical symptoms and signs
to assess the clinical effectiveness of biliary drainage and the occurrence of AEs
after endobiliary RFA. If clinically successful biliary drainage was not achieved,
additional drainage procedures, including percutaneous transhepatic biliary drainage,
were considered. At 1-month intervals, patients were followed up and subjected to
blood tests and simple abdominal X-rays. Computed tomography was performed as needed
for suspected recurrent biliary obstruction (RBO) or AEs, or every 2 months to assess
disease progression in patients who received chemotherapy. Chemotherapy using gemcitabine
and cisplatin was administered according to patient performance status. Scheduled
secondary interventions including regular stent exchange or additional RFAs after
RBO were not performed for both groups.
Definitions and outcome measurements
The primary outcome was the cumulative time to RBO and secondary outcomes were the
technical success rate, clinical success rate, AE rate, overall survival (OS), and
factors affecting time to RBO.
RBO was defined as a composite outcome of stent occlusion or migration, which was
confirmed by clinical signs such as jaundice and/or cholangitis and imaging findings.
[13 ] The cumulative time to RBO was defined as the time from the date of stent placement
until the date of RBO, death, loss to follow-up, or survival at the end of the study
with no occlusion. Technical success was defined as successful completion of stent
placement after endobiliary RFA or successful completion of stenting only. The RFA
session was considered successful if the ablative energy was delivered for the scheduled
time (90 to 120 seconds) with coverage of all tumor lesions. Clinical success was
defined as resolution of symptoms, including obstructive jaundice and cholangitis,
without further interventions within 1 week [14 ]
[15 ]. AEs were assessed during follow-up after the initial intervention, with the exception
of RBO, and were evaluated according to the grading system of the American Society
for Gastrointestinal Endoscopy [16 ]. OS was defined as the time from initial stenting to death or the day of the last
follow-up. Factors affecting time to RBO—including baseline characteristics, RFA treatment,
stent type, and chemotherapy status—were evaluated.
Statistical analysis
Propensity score matching was performed to reduce the effects of potential confounding
bias on measured outcomes because of differences between the groups in patient baseline
characteristics. Bismuth classification, ECOG performance status, tumor stage, and
stent type were selected as the observed covariates. The aforementioned variables
were selected based on analysis of previous literature [17 ]
[18 ]. Propensity scores were estimated using logistic regression models based on this
set of covariates. The groups were matched by using 1:1 optimal matching without replacement
and a caliper of width of 0.2 [19 ].
Continuous variables are presented as the median and interquartile range (IQR), and
categorical variables are presented as the frequency and percentage. Continuous variables
were compared using Student’s t -test or the Wilcoxon rank-sum test, and categorical variables were compared using
the Pearson chi-squared test or Fisher’s exact test. The median cumulative time to
RBO and OS with 95% confidence intervals (CIs) were estimated using the Kaplan-Meier
method and compared with the log-rank test. Data were censored for patients who died,
were lost to follow-up, or did not develop RBO by the end of the study period. A Cox
proportional hazard model was used to identify factors affecting the time to RBO.
Statistical analysis was conducted using SPSS version 21.0 for Windows software (IBM
Corp., Armonk, New York, United States), and P <0.05 was considered indicative of statistical significance.
Results
One hundred twenty-nine patients with inoperable hilar CCA were included in the analysis;
34 patients received RFA and 95 patients did not receive RFA. After 1:1 propensity
score matching, 32 patients each were categorized in RFA + stent and stent-only groups
([Fig. 5 ]). No significant differences in Bismuth classification, ECOG status, tumor stage,
or stent type were detected between the two groups ([Table 1 ]).
Fig. 5 Study flow chart. RFA, radiofrequency ablation.
Table 1 Baseline characteristics.
Characteristics
Before propensity score matching
After propensity score matching
RFA + stent group
Stent-only group
SMD
RFA + stent group
Stent-only group
SMD
Values are n (%) or median (interquartile range).
Age, years
76 (68–78)
73 (65–80)
75 (67–78)
74 (69–79)
Sex, female
22 (64.7)
44 (50.6)
20 (62.5)
17 (53.1)
Bismuth classification
0.298
0.025
12 (35.3)
53 (55.8)
11 (34.4)
12 (37.5)
22 (64.7)
42 (44.2)
21 (65.6)
20 (62.5)
ECOG performance status
0.271
< 0.001
24 (70.6)
48 (50.5)
25 (78.1)
22 (68.8)
10 (29.4)
47 (49.5)
7 (21.9)
10 (31.3)
Stage
0.129
< 0.001
21 (61.8)
51 (53.7)
21 (65.6)
19 (59.4)
13 (38.2)
49 (51.6)
11 (34.4)
13 (40.6)
Types of deployed stent
0.317
0.066
15 (44.1)
52 (54.7)
15 (47.3)
16 (50.0)
19 (55.9)
43 (45.3)
17 (53.1)
16 (50.0)
Total bilirubin, mg/dL
6.3 (4.4–11.4)
6.3 (4.3–9.3)
6.3 (4.5–13.0)
6.4 (6.1–8.3)
Chemotherapy, n (%)
18 (52.9)
36 (37.9)
17 (53.1)
15 (46.9)
[Table 2 ] lists the outcomes of both groups after propensity score matching. Technical success
was achieved in all patients. The clinical success rate was 93.8% (30 of 32) in the
RFA + stent group and 87.5% (28 of 32) in the stent-only group (P= 0.672). The RBO rate was 75.0% (24 of 32) in the RFA + stent group and 84.4% (27 of
32) in the stent-only group (P =0.351). The overall AE rates in the RFA + stent and stent-only groups were 12.5%
and 9.4% (P= 1.000), respectively. In both groups, cholangitis (RFA + stent group, n=3; stent-only
group, n=2) and cholecystitis (stent-only group, n=1) occurred within 30 days of the
RFA procedure and were graded as mild and treated conservatively without additional
interventions. In the RFA + stent group, one patient experienced hemobilia after PS
removal 182 days after RFA, which was successfully controlled without transfusion
by immediate insertion of a SEMS into the bile duct.
Table 2 Comparison of outcomes between both groups.
Characteristics
RFA + stent group
Stent-only group
P value
Values are n (%).
Technical success
32 (100.0)
32 (100.0)
1.000
Clinical success
30 (93.8)
28 (87.5)
0.672
Recurrent biliary obstruction
24 (75.0)
27 (84.4)
0.351
Overall AEs, n (%)
4 (12.5)
3 (9.4)
1.000
0 (0.0)
0 (0.0)
1.000
3 (9.4)
2 (6.3)
1.000
0 (0.0)
1 (3.1)
1.000
1 (3.1)
0 (0.0)
1.000
Perforation
0 (0.0)
0 (0.0)
1.000
In the RFA + stent group, the number of RFA applications required to ablate all lesions
was one in 21.9% (7 of 32), two in 31.3% (10 of 32), and three in 46.9% (15 of 32)
of the patients. The median total ablation length was 33 mm (IQR, 22–40) and the median
total ablation time was 240 seconds (IQR, 120–330). The lengths of the electrodes
used were 11 mm (55.6%), 18 mm (31.9%), and 22 mm (12.5%) ([Table 3 ]).
Table 3 Outcomes of endobiliary RFA.
Characteristics
n=32
Values are n (%), n/N (%), or median (interquartile range).
Total number of RFA applications
7 (21.9)
10 (31.3)
15 (46.9)
Total ablation length, mm
33 (22–40)
Total ablation time, sec
240 (120–330)
Electrode lengths used, mm
40/72 (55.6)
23/72 (31.9)
9/72 (12.5)
The median cumulative time to RBO was 242 days (95% CI 181–309 days) in the RFA +
stent group and 168 days (95% CI 159–281 days) in the stent-only group (P= 0.031). The median OS was 337 days (95% CI 252–404 days) in the RFA + stent group
and 296 days (95% CI 289–383 days) in the stent-only group (P= 0.260) ([Fig. 6 ]).
Fig. 6
a Kaplan-Meier curves of cumulative time to recurrent biliary obstruction (RBO). The
median time to RBO was significantly longer in the radiofrequency ablation (RFA) +
stent group than in the stent-only group (242 days vs 168 days, P = 0.031). b Kaplan-Meier curves of overall survival (OS). There was no significant difference
in median OS (337 days vs 296 days, P = 0.260) between the two groups.
Discussion
Few studies have evaluated endobiliary RFA for hilar CCA due to the complexity of
the hilar strictures and reports of severe AEs such as hemobilia, pseudoaneurysm,
and hepatic infarction due to arterial thrombosis [4 ]
[5 ]
[20 ]. Compared with the distal bile duct, the curved structures with side branches of
the hilar bile duct can reduce the uniformity of the therapeutic effect of RFA, and
its thin fibromuscular/serosal layers and location adjacent to the hepatic artery
and portal vein increase the risk of severe AEs [21 ]
[22 ].
A new RFA system that can maintain a target temperature has been introduced to improve
efficacy and reduce serious AEs [6 ]
[17 ]
[23 ]. Temperatures >100°C not only increase the risk of AEs but also cause the formation
of a “coagulum,” which increases resistance to alternating current and reduces RFA
effectiveness. The new RFA system is safe for use in the hilar bile duct because it
monitors the temperature at the catheter tip and maintains a preset temperature by
means of an automatic on-and-off function [6 ]
[7 ].
In this study, we evaluated the effectiveness and safety of temperature-controlled
endobiliary RFA followed by stent placement compared with stent placement only for
the palliative treatment of inoperable hilar CCA. In contrast to the recent RCTs by
Albers et al. [24 ] and Jarosova et al. [25 ], our results showed a significant improvement in cumulative time to RBO in the RFA
+ stent group compared with the stent-only group (242 days vs 168 days, P = 0.031). We speculate that the benefit of RFA on the median time to RBO is likely due
to the study design; unlike the aforementioned study in which the majority of patients
had pancreatic cancer, we recruited and evaluated only patients with pathologically
confirmed CCA. Given that the primary mechanism of obstruction by pancreatic cancer
is compression rather than wall infiltration, ablation treatment may be more effective
in reducing the tumor load in CCA because the latter typically grows along the bile
duct wall and is of limited thickness [26 ].
Findings about the correlation between endobiliary RFA and increased survival are
conflicting [3 ]
[8 ]
[27 ]. In this study, there was a marginal trend toward increased OS in the RFA + stent
group compared with the stent-only group (337 days vs 296 days, P= 0.260). Because endobiliary RFA is a local ablative treatment that delivers thermal
energy to accessible tissues, it would have been difficult to demonstrate a significant
OS improvement in our population, which included a large number of CCA patients with
distant metastasis (34.4% in the RFA + stent group and 40.6% in the stent-only group).
Although the trend toward an improved OS in the RFA + stent group suggests systemic
effects of RFA on tumor-specific cytotoxic T-cell responses, modulation of circulating
immune cells and cytokines, and amelioration of immunosuppression, further research
is needed [28 ].
In the meta-analysis by Sofi et al. [29 ], two deaths as well as mild-to-moderate cholangitis and acute cholecystitis were
reported among 239 patients who underwent endobiliary RFA; severe AEs were associated
with injury to the major vessels adjacent to the bile ducts, which may have been caused
by excessive heat during RFA [30 ]. In this study, all AEs were mild to moderate, and no severe AE was attributed to
endobiliary RFA. In the RFA + stent group, cholangitis (n=3), which was probably caused
by poor drainage of the bile duct due to tumor necrosis after RFA, improved after
conservative treatment. Although there was one case of hemobilia caused by the separation
of deep necrosis from the perihilar tissue after PS removal in the RFA + stent group,
it was stopped by immediate insertion of a SEMS into the bile duct. The comparable
rates of AEs in the RFA + stent and stent-only groups (12.5% vs 9.4%; P= 1.000) indicate the safety of temperature-controlled RFA in patients with hilar CCA.
The favorable results of this study may be due to the variety of electrode types used
and the temperature-control function of the RFA system [31 ]. The most widely used Habib EndoHBP (Boston Scientific, Marlborough, Massachusetts,
United States) is a 180-cm-long 8F bipolar catheter with two 8-mm electrodes spaced
8 mm apart for an expected ablation length of 24 mm; it has a long ablation zone,
which may not be suitable for use in the hilar bile duct. The RF catheter used in
this study (ELRA; STARmed) allows the use of short electrodes (11 or 18 mm), thereby
promoting effective RFA in the hilar bile duct. This system allows continuous maintenance
of the selected electrode temperature during RFA, ensuring safe and effective RFA
procedures even in the presence of complex hilar bile duct structures.
We aimed to improve the clinical outcomes of RFA. First, we measured stricture thickness
using IDUS to reduce the risk of AEs in all patients. Second, we attempted to ablate
all lesions using a short catheter (11 mm, 55.6%; 18 mm, 31.9%) by dividing strictures
into several sections rather than ablating a single long segment. Of the patients,
46.9% (15 of 32) underwent three RFA procedures. Fractionated applications of RFA
with a short catheter may have maximized intimate contact with the target tumor burden
even in the angulated hilar bile duct. Third, we adjusted the ablation time of the
right hepatic duct to 90 seconds, based on the risk of severe bleeding due to complex
duct–vascular contact of the right hepatic duct and the right hepatic artery. Because
the right hepatic artery typically crosses the posterior aspect of the CHD and then
runs along the right hepatic duct, RFA of the right hepatic duct requires caution.
Finally, subsequent stent placement for biliary drainage was performed immediately
following endobiliary RFA to maintain bile duct patency. Ablation using a short RF
catheter (11 or 18 mm) for 90 to 120 seconds at 7 W of power and a temperature of
80°C was effective and safe in patients with hilar CCA.
This study had several limitations. First, although propensity score matching was
adjusted for background characteristics, the small sample size reduced the reliability
of the statistical analysis and hampered subgroup analyses. Second, the type of deployed
stent was not controlled for, although the types were distributed similarly between
the groups. It was difficult to use the same type of stent for drainage due to differences
in patient life expectancy. Third, there was no direct comparison of available RFA
systems, so our results do not confirm that temperature-controlled RFA has better
outcomes than power-controlled RFA. Further studies of the clinical outcomes of such
systems are needed. Fourth, because RFA was performed in a single session, we could
not determine the optimal frequency and interval. Fifth, biologic tests of antitumor
cells and immunity were not performed, so no conclusions could be drawn about the
systemic effects of RFA. The effect of endobiliary RFA on the systemic inflammatory
response needs to be evaluated. Finally, the procedures were performed by highly skilled
endoscopists, so the results may not be directly applicable to low-volume centers.
Conclusions
In conclusion, temperature-controlled endobiliary RFA is an effective and safe palliative
treatment that provides sufficient destruction of local tumor without increasing the
risks of severe AEs, thereby improving the effectiveness of subsequent treatments
in patients with inoperable hilar CCA.
