Keywords RFA and ablative methods - Strictures - Endoscopy Upper GI Tract - Pancreatobiliary
(ERCP/PTCD)
Introduction
Early diagnosis of papillary neoplasia is challenging because the symptoms usually
appear in cases of advanced carcinoma [1 ]. Most cases are diagnosed incidentally during endoscopy for other indications. In
addition, endoscopic biopsies are mandatory for histologic confirmation of adenoma
before the therapeutic approach [2 ]
[3 ].
Surgery is considered the gold standard procedure for therapeutic resection. However,
endoscopy can be considered in selected cases because as it is a less invasive approach
[2 ]
[3 ]. European Society of Gastrointestinal Endoscopy guidelines recommend endoscopic
papillectomy (EP) in ampullary adenoma without intraductal extension but suggest considering
surgical treatment when the endoscopic procedure is not feasible (size > 40 mm and
intraductal involvement > 20 mm) [1 ].
Despite the effectiveness of endoscopic resection and the lower morbidity and mortality
compared with pancreatoduodenectomy, it determines recurrence in about 30% of cases
[4 ]
[5 ]
[6 ]
[7 ]
[8 ]
[9 ]. Given the recurrence rate of endoscopic resection and the risks related to surgery,
recent studies have shown the benefits of radiofrequency ablation (RFA) for residual
lesions and as a complementary therapy for an intraductal extension [2 ]
[4 ]
[5 ]
[10 ]
[11 ]
[12 ]
[13 ]
[14 ]
[15 ].
RFA acts directly on residual neoplastic tissue, causing necrosis from the resulting
thermal energy, and determines highly immunogenic intracellular components like heat
shock proteins [16 ]
[17 ]
[18 ]. To better understand the outcomes of this novel approach, we performed a systematic
review and meta-analysis evaluating the efficacy and safety of RFA for residual or
recurrent lesions with intraductal extension after EP.
Patients and methods
Protocol and registration
The study was registered in the International Prospective Register of Systematic Reviews
(PROSPERO) with the file number (CRD42023395394). This review and meta-analysis were
performed under the recommendations of the Cochrane Handbook of Systematic Reviews
of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-analysis
(PRISMA) Guidelines [19 ].
Eligibility criteria
The inclusion criteria included clinical trials, cohort studies, and case series that
investigate patients with residual or recurrent lesions extending into the pancreatic
or biliary duct after EP treated with RFA. Exclusion criteria included case reports,
insufficient data, studies from the same authors that had been updated, and follow-up
period of fewer than 10 months.
Search and study selection
The studies were identified through a search in electronic databases (MEDLINE, Embase,
Cochrane), from inception until October 20, 2023. No date or language restrictions
were set. Two reviewers achieved the selection of studies independently, and a third
reviewer was consulted in cases of disagreement. The following search strategy was
used for the MEDLINE database: (Papillary Adenoma OR Adenomas, Bile Duct OR Ampulla
of Vater OR Hepatopancreatic Ampulla OR Major Duodenal Papilla OR Bile Duct) AND (Radiofrequency
Catheter Ablation OR Electrical Catheter Ablation OR Catheter Ablation OR Radiofrequency
OR Ablation Techniques OR Radiofrequency Therapy OR Electrocoagulation OR Electrocautery
OR Thermocoagulation)’.
Data collection process
Data extraction was done by filling out a spreadsheet. The following data were extracted:
name and year of the study, number of patients undergoing EP, number of patients undergoing
RFA, recurrence rate for evaluation of clinical success, surgical conversion rate,
number of adverse events (AEs), including cholangitis, perforation, stenosis, pancreatitis,
bleeding.
Risk of bias and quality of studies
For the analysis of the validity, reliability, and relevance of studies, two independent
reviewers assessed the risk of bias using the Joanna Briggs Institute Critical Appraisal
Tool (https://jbi.global/critical-appraisal-tools ) ([Table 1 ]), a specific tool for case series that evaluates the following items: patient demographic
characteristics, patient history, current clinical condition on presentation, diagnostic
tests or assessment methods and their results, intervention(s) or treatment procedure(s),
post-intervention clinical condition, AEs (harms) or unanticipated events, and takeaway
lessons. Additionally, a tool from the Robvis website was employed to create a table
summarizing the risk of bias analysis (https://www.riskofbias.info/welcome/robvis-visualization-tool ). The risk of bias was graduated in low, high or very high risk.
Table 1 JBI tool for risk of bias assessment.
Study
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
Overall
D1: Inclusion criteria
D2: Condition evaluation
D3: Condition identification
D4: Consecutive inclusion
D5: Complete inclusion
D6: Study demographic report
D7: Clinical information
D8: Outcomes and follow-up
D9: Site demographic information
D10: Statistical analysis
Cho et al.
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Include
Dahel et al.
Yes
Unclear
Yes
Yes
Yes
No
Yes
Yes
No
Unclear
Include
Trigali et al.
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Include
Choi et al.
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Include
Bruwier et al.
Yes
Unclear
Unclear
Yes
Yes
No
Yes
Yes
No
Unclear
Include
Camus et al.
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Include
Rustagi et al.
Unclear
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Unclear
Include
The quality of evidence was appraised using the Grading of Recommendations Assessment,
Development, and Evaluation (GRADE) system from the GRADEpro - Guideline Development
Tool software (McMaster University, Ontario, Canada). This system considers the following
items: design, risk of bias, precision, indirect evidence, inconsistency, publication
bias, effect magnitude, dose dependence, and confounding bias ([Table 2 ]). The quality of evidence was graded as high, moderate, low, or very low [20 ].
Table 2 Quality of evidence by Grading of Recommendations Assessment, Development and Evaluation
(GRADE) guidelines [20 ].
Certainty assessment
No. of patients
Effect
Certainty
№ of studies
Study design
Risk of bias
Inconsistency
Indirectness
Imprecision
Other considerations
Radioablation
Relative
Absolute
7
Observational studies
Not serious
Not serious
Not serious
Not serious
None
50/124 (40.3%)
–
–
–
⨁⨁◯◯
7
Observational studies
Not serious
Not serious
Not serious
Not serious
None
95/124 (76.6%)
–
–
–
⨁⨁◯◯
7
Observational studies
Not serious
Not serious
Not serious
Not serious
None
3/124 (2.4%)
–
–
–
⨁⨁◯◯
7
Observational studies
Not serious
Not serious
Not serious
Not serious
None
4/124 (3.2%)
–
–
–
⨁⨁◯◯
7
Observational studies
Not serious
Not serious
Not serious
Not serious
None
15/124 (12.1%)
–
–
–
⨁⨁◯◯
7
Observational studies
Not serious
Not serious
Not serious
Not serious
None
28/124 (22.6%)
–
–
–
⨁⨁◯◯
7
Observational studies
Not serious
Not serious
Not serious
Not serious
None
6/124 (4.8%)
–
–
–
⨁⨁◯◯
Outcomes and definitions
Outcomes evaluated were the clinical success, defined as the rate of patients who
did not experience a recurrence during follow-up, surgical conversion rate, recurrence,
and number of AEs such as biliary stenosis, pancreatitis, and cholangitis.
Data synthesis and statistical analysis
For continuous variables, the mean difference and standard deviation were calculated
using inverse variance. For dichotomous variables, the risk difference (RD) was calculated
using Mantel–Haenszel, along with a corresponding 95% confidence interval (CI). When
the variance was expressed as a range, the mean and variance of the sample were estimated
using the Hozo test [21 ]. Comprehensive Meta-Analysis V4 software was utilized for data analysis, forest
plot generation, and CI calculation [22 ]. Data heterogeneity was assessed and quantified according to the Higgins Method
(I2 ). Pooled estimates and the 95% CIs were calculated using a random-effects model.
Radiofrequency ablation
RFA was conducted after EP in all patients who exhibited residual or recurrent lesions.
ID-RFA was conducted using RFA catheters (ELRA; STARmed, Goyang, Korea) or (Habib
EndoHPB, Boston Scientific, London, U.K). The RFA catheters were inserted into the
biliary or pancreatic duct through 0.025- or 0.035-inch guidewires.
The ELRA catheter had a diameter of 7F and a length of 175 cm, equipped with bipolar
probes consisting of electrodes of various lengths (11 mm, 18 mm, 22 mm, and 33 mm),
employed to accommodate diverse anatomical and geometric variations at the target
ablation site. The VIVA Combo generator (STARmed, Seoul, South Korea) was employed
for intraductal RFA delivery, providing precise control over power settings, target
temperature, and impedance [14 ]
[15 ]
[23 ].
The Habib catheter it is an 8F (2.7 mm) sizable bipolar RFA probe, extending 180 cm
in length, and is equipped with two ring electrodes that are spaced 8 mm apart, and
the distal electrode is positioned 5 mm from the front edge. The catheter was attached
to an electrosurgical generator, with options including the RITA 1500X from Angiodynamics
in Latham, New York, United States the Erbe system from Surgical Technology Group
in Hampshire, England, UK, or the Beamer from ConMed [12 ]
[13 ].
Results
Result of searches and characteristics of the included studies
The initial search found a total of 4,546 studies. After removing duplicate articles
and
reviewing titles and abstracts, 20 case series were found eligible for full-text analysis.
We excluded eight case reports ([Fig. 1 ]). A total of 12 were utilized for qualitative synthesis and seven for quantitative
synthesis, totaling 124 patients [12 ]
[13 ]
[14 ]
[15 ]
[23 ]
[24 ]
[25 ]. Five studies were excluded from the quantitative analysis due to duplicate data
([Table 3 ]).
Fig. 1 PRISMA flow chart [19 ].
Table 3 Summary of the included studies.
Study
Study Design
Age (Mean)
Number of Patients (RFA)
Neoplasia
Devices
Duration (RFA)
Power Setting (RFA)
Number of Sessions (mean)
Follow-up (mean)
CS, case series; HGD, high-grade dysplasia; LGD, low-grade dysplasia; Tis, carcinoma
in-situ; ADC, adenocarcinoma; IMC, intramucosal carcinoma; mo, months; NM, not mentioned;
s, seconds; CBD, common bile duct; PD, pancreatic duct; NET, neuroendocrine tumor.
Cho et al. 2023 [15 ]
Prospective series
61.2
29
21 (LGD);
ELRA (STARmed)
120s (CBD) 30 (PD)
7w
1.5
25 mo
Dahel et al. 2023 [24 ]
Retrospective series
NM
25
10 (LGD);
NM
NM
NM
1.3
36 mo
Trigali et al. 2021 [14 ]
Prospective series
73
9
4 (LGD);
ELRA (STARmed)
120s
10w
1.6
26.2 mo
Choi et al. 2021 [23 ]
Retrospective series
56.7
10
8 (LGD);
ELRA (STARmed)
65s (CBD) 15s (PD)
7w
1
10 mo
Bruwier et al. 2020 [25 ]
Prospective series
73
17
14 (LGD);
ELRA (STARmed)
30–240s
7–10w
1.8
12 mo
Camus et al. 2018 [12 ]
Prospective series
67
20
15 (LGD);
Habib (Boston)
30s
10w
1
12 mo
Rustagi et al. 2016 [13 ]
Retrospective series
68
14
08 (LGD);
Habib (Boston)
90s
7–10w
1.6
16 mo
Clinical success
All included studies assessed clinical success. RFA after EP revealed a clinical success
rate of 75.7% (95% CI 65.0–88.0%; I2 = 23.484) ([Fig. 2 ]).
Fig. 2 Forest plot for rate of clinical success, total adverse rates, surgical conversion
rate and recurrence using the random-effect model. CI, confidence interval.
Recurrence
All studies reported this outcome. The meta-analysis resulted in a recurrence of 24.3%
(95% CI 16.0–35.0%; I2 < 0.001) ([Fig. 2 ]).
Surgical conversion rate
All studies reported this outcome. The meta-analysis resulted in a surgical conversion
rate of 6.7% (95% CI 3.2–13.4%; I2 < 0.001) ([Fig. 2 ]).
Total adverse events
All included studies reported the rate of AEs during the follow-up period. The rate
of total AEs of 41.1% (95% CI 30.7–52.4%; I2 = 27.541) ([Fig. 2 ]).
Biliary stricture
All included studies reported the incidence of RFA-related biliary stricture. The
incidence of biliary stricture was 22.2% (95% CI 12.1–28.4%; I2 = 61.030) ([Fig. 3 ]).
Fig. 3 Forest plot for rate of adverse events, using the random-effect model. CI, confidence
interval.
Pancreatitis
All included studies reported the incidence of RFA-related pancreatitis. The incidence
of pancreatitis was 14.3% (95% CI 8.8–22.3%; I2 < 0.001) ([Fig. 3 ]).
Cholangitis
All included studies reported the incidence of RFA-related cholangitis. The incidence
of cholangitis was 7.0% (95% CI 3.3–14.5%; I2 < 0.001) ([Fig. 3 ]).
Bleeding
All included studies reported the incidence of RFA-related bleeding. The incidence
of bleeding was 4.0% (95% CI 1.7–9.3%; I2 < 0.001) ([Fig. 3 ]).
Perforation
No perforations were related to endoscopic resection and RFA in any of the evaluated
studies.
Discussion
This was the first systematic review and meta-analysis evaluating the outcomes of
RFA for residual lesions after EP, showing that this technique may be effective in
managing this challenging condition but with a very high rate of AEs.
This meta-analysis revealed a high clinical success rate; however, this should be
evaluated cautiously due to the short follow-up period of patients and the heterogeneity
of the sample. The minimum follow-up period of 10 months and the maximum of 36 months
do not allow for an adequate assessment of the recurrence rate. Recent data suggest
that recurrence can occur even after 5 years, therefore, follow-up for this period
is necessary to assess the recurrence rate properly [26 ]
[27 ]. The heterogeneity of our meta-analysis is demonstrated by including patients with
intramucosal ADCs in some studies because adenomas have a lower recurrence rate than
ADCs. Furthermore, the included ADCs were not classified by their histological type,
and it is well known that the pancreaticobiliary-type is more aggressive than the
intestinal-type. Pancreaticobiliary-type and other undifferentiated cancers have a
high capacity for local dissemination and a high recurrence rate, deserving a multidisciplinary
approach to management [3 ]
[28 ]
[29 ]. All guidelines recommend referring the patient for surgery in case of papillary
adenocarcinoma [1 ]
[2 ]
[3 ]. However, some authors advocate less invasive procedures for early-stage adenocarcinoma,
and it is essential to differentiate Tis carcinoma, which does not invade the lamina
propria and is associated with a lower incidence of lymph node invasion, and T1a carcinoma,
which invades a lamina propria and is associated with more than 20% lymph node invasion
[1 ]
[3 ]
[30 ]. ESGE recommends that EP for Tis ampullary cancer might be considered sufficient
when there is no residual disease [1 ]. Thus, some studies have reported that EP alone may achieve curative resection in
cases of Tis and T1a carcinoma without lymphatic invasion [30 ]
[31 ]
[32 ]. Moreover, despite the absence of studies evaluating the use of RFA for neuroendocrine
tumors (NET), the study published by Dahel et al. included a single patient with this
neoplasm [24 ]. There are no available studies assessing its use for duodenal NET; however, two
meta-analyses published in 2023 demonstrated positive outcomes in the use of RFA for
pancreatic NET [33 ]
[34 ].
In addition, the rate of AEs was higher than evidenced in studies that analyze RFA
for malignant biliary strictures [35 ]
[36 ]. The most common AE was biliary stenosis, but we could not evaluate the correlation
with the absence of a prophylactic biliary stent. ESGE suggests using a temporary
biliary stent with a complementary technique, such as RFA for ampullary adenoma with
≤ 20 mm intraductal extension. The Expert Consensus mentioned that stent placement
in case of residual tissue after EP can facilitate the inspection of the distal common
bile duct, but no consensus was achieved about this matter [2 ]. In addition, ablation with higher power and longer time may be associated with
a higher incidence of biliary stricture. Most studies have applied energy of 7 to
10 W for 90 to 120 seconds for each intrabiliary RFA application. Although, it was
also not possible to evaluate this correlation based on the data available, further
research can identify the optimal settings for these parameters for treating ampullary
adenomas [15 ]
[23 ]
[37 ].
In this meta-analysis, the second most significant AEs was pancreatitis. Unfortunately,
it was not possible to classify the severity of the AEs evaluated due to the scarcity
of data provided. In the updated ESGE Guideline on ERCP-related AEs, pancreatic duct
stenting, rectal nonsteroidal anti-inflammatory drugs, and high-volume hydration were
recommended to prevent post-ERCP pancreatitis [1 ]
[38 ]. These recommendations can also be applied to patients after EP to decrease the
risk of post-ERCP pancreatitis. In a subgroup analysis, including three studies involved
in the meta-analysis, there were six cases (13%) of pancreatitis among the 44 patients
who used prophylactic stents, representing a significant rate of events. However,
it was not possible to carry out a comparative analysis with the group of patients
who did not use a stent due to the lack of data.
In addition to the limitations already discussed, our study has other relevant limitations.
The most important is the small number of studies and patients included in the analysis.
However, the reason for that is the lack of large studies on this subject, and we
performed the analysis with the available data. Also, as it is an approach that has
emerged in recent years, no randomized clinical trials and cohort studies are available,
contributing to the high risk of bias in all the included studies. Furthermore, while
some studies conduct RFA for patients with residual lesions shortly after papillectomy,
others address patients with either residual or recurrent lesions. However, the lack
of standardization in defining recurrence across these studies presents another limitation.
This inconsistency impedes a thorough assessment of RFA efficacy for each specific
situation separately. Relevant data such as the number of radiofrequency sessions
performed on each patient, the use of combined therapy involving argon plasma coagulation,
and the correlation between the type of stent and incidence of pancreatitis or bile
duct narrowing were only reported in some studies, which precludes a more detailed
analysis.
In summary, this study showed that using RFA for residual lesions after EP has a significant
clinical success rate, although it reveals a high rate of AEs. These events may be
associated with factors such as the absence of prophylactic biliary or pancreatic
stents. With our results, we believe this method may become the gold standard technique
to avoid complex surgeries with a high rate of complications, such as pancreaticoduodenectomy.
Despite the high rate of AEs revealed in our meta-analysis, most of them were mild
and self-limited, and they become less relevant when comparing surgery-related complications.
Conclusions
RFA is feasible and appears to be effective for managing residual lesions after EP.
However, long-term follow-up and high-quality studies are required to confirm our
findings. In
addition, to improve safety before disseminating this therapy, we should carefully
assess the
high rate of AEs related to RFA after EP.
