Tools and Techniques of Endoscopic Radiofrequency Ablation in Pancreaticobiliary Diseases: A Narrative Review

Abstract

Background

Endoscopic radiofrequency ablation (RFA) has become an established minimally invasive option for patients with pancreaticobiliary tumors who are not candidates for resection or who have obstructing lesions of the bile duct or pancreas. This narrative review summarizes the devices, technique, and supporting clinical evidence for endoscopic RFA, organized by clinical indication rather than by device alone.

Objective

RFA delivers high-frequency alternating current that produces localized thermal injury and coagulative necrosis. Two endoscopic delivery routes are in use: an ERCP-guided approach, which treats intraductal lesions such as malignant biliary strictures and ampullary tumors using bipolar endobiliary catheters, and an EUS-guided approach, which uses a monopolar needle electrode to ablate pancreatic parenchymal lesions, including solid tumors, neuroendocrine tumors, and pancreatic cysts and intraductal papillary mucinous neoplasms (IPMN). Hybrid needles that combine fine-needle aspiration with ablation are an emerging, still investigational, third category.

Synthesis/Results

Across the published series, technical success generally exceeds 90% for both approaches, with adverse event rates that are mostly low and dominated by self-limited abdominal pain, mild pancreatitis, or cholangitis; perforation and clinically significant bleeding are uncommon when established energy and safety-margin parameters are followed. A proposed, but still poorly characterized, immune-stimulating effect of ablation-induced tissue injury is discussed as a hypothesis rather than an established mechanism.

Conclusions

Taken together, the available data support endoscopic RFA as a useful adjunct in the multidisciplinary management of pancreaticobiliary disease, but most of the supporting literature consists of small, single-arm, or retrospective series. Prospective, ideally randomized, studies with standardized energy protocols and longer follow-up are needed before the technique’s role in routine practice, particularly for pancreatic cystic lesions and IPMN, can be considered settled.

Introduction

Radiofrequency ablation was first established as a percutaneous treatment for hepatic tumors and was later adapted for esophageal dysplasia before being applied, via endoscope, to the pancreas and bile duct. In the pancreaticobiliary setting it is used mainly in patients whose tumors are unresectable or who need palliation of obstructing disease.

Pancreatic cancer is the third leading cause of cancer death in the United States, with an estimated 67,440 new cases and 51,750 deaths projected for 2025 [1]. Five-year survival across all stages remains poor, at approximately 13%, because most patients present with locally advanced or metastatic disease [1]. Pancreatic ductal adenocarcinoma (PDAC) accounts for roughly 85–90% of pancreatic malignancies, with neuroendocrine tumors, cystic neoplasms, and acinar cell carcinoma comprising most of the remainder [1]. Biliary tract cancers, principally cholangiocarcinoma, are considerably less common, an estimated 8,000 new cases annually in the United States compared with more than 60,000 for pancreatic cancer, but carry a similarly poor prognosis and are frequently diagnosed late because early symptoms are absent or nonspecific [2, 3].

The rising number of patients presenting with unresectable pancreaticobiliary cancer has increased interest in effective palliative options, and RFA is one of these. It works by delivering high-frequency alternating current, typically in the 450–500 kHz range, to generate localized heat and coagulative necrosis in the target tissue [4]. Two endoscopic approaches are used clinically:

• ERCP-guided RFA, in which a flexible endoscope is combined with fluoroscopy and cholangiography to navigate the pancreaticobiliary ducts and ablate malignant biliary strictures, intraductal lesions, and ampullary tumors [5].

• EUS-guided RFA, in which a linear-array echoendoscope provides real-time imaging of the pancreas, liver, and bile ducts, allowing thermal energy to be delivered to deeper parenchymal lesions while sparing surrounding tissue. In addition to direct cytotoxicity, ablation-induced tissue injury has been proposed to release tumor-associated antigens and so trigger a secondary immune response; this remains a hypothesis drawn mainly from preclinical and hepatic-tumor data rather than a mechanism established in the pancreaticobiliary literature, and it is not yet a basis for clinical decision-making [6].

This review summarizes the devices and technique used for endoscopic RFA (Fig. 1), then presents the supporting clinical evidence organized by indication, including a discussion of EUS-guided ablation for pancreatic cystic lesions and IPMN, which is increasingly cited as a target for this technology but has not been well covered in prior reviews.

Fig. 1.

Endoscopic RFA approaches, indications, and mechanism

Methods

This is a narrative, rather than systematic, review. Relevant articles were identified through PubMed/MEDLINE searches combining the terms 'radiofrequency ablation,' 'endoscopic ultrasound,' 'ERCP,' 'biliary obstruction,' 'cholangiocarcinoma,' 'pancreatic neuroendocrine tumor,' and 'intraductal papillary mucinous neoplasm,' supplemented by manual review of reference lists from identified articles and recent society technology reviews. Case reports with fewer than five patients and non-English-language articles without an available translation were excluded. Because the underlying evidence base is composed largely of single-arm and retrospective studies, no formal quality scoring or meta-analytic synthesis was attempted; findings are summarized descriptively by clinical indication.

Tools and Technique

ERCP-Guided Devices

ERCP was originally a diagnostic procedure and has since become a therapeutic platform for relieving obstruction of the hepatic, pancreatic, and biliary ducts caused by stones, tumors, or strictures. Stenting remains the mainstay of palliation for malignant biliary obstruction, but adding RFA at the time of stenting is increasingly used as a well-tolerated adjunct [7]. Two bipolar catheter systems are in clinical use.

Habib EndoHPB (Boston Scientific, USA). The Habib EndoHPB was the first RFA catheter approved in the United States for ablation of malignant or benign tissue in the pancreaticobiliary tract. It is a bipolar device, 7 Fr (2.3 mm) in diameter and 1.8 m long, compatible with endoscope working channels of 3.2 mm or less, and it connects to standard electrosurgical generators through an adaptor cable, which limits additional equipment requirements. The catheter is advanced over a 0.035-inch guidewire into the target duct under fluoroscopic guidance, with the most distal tissue treated first and a sequential withdrawal technique used for longer strictures. Recommended settings are 7 W above and 10 W below the biliary bifurcation, applied for 90–120 s with a 60-second pause before repositioning [4].

ELRA Endobiliary RFA Catheter (Taewoong Medical, South Korea). The ELRA catheter uses a bipolar, radially expandable electrode that conforms to the duct lumen, allowing more uniform circumferential ablation than a fixed-diameter catheter. It is single-use, available in 8, 10, and 12 mm expanded diameters, and incorporates a temperature sensor that feeds back to the generator (the VIVA Combo RF generator) to limit excess thermal injury. After standard ERCP cannulation and guidewire placement across the stricture, the catheter is advanced to the target and energy is delivered at a target temperature of approximately 80 °C and a power of 7–10 W, with each application lasting 90–120 s depending on lesion size.

EUS-Guided Devices

EUSRA (Taewoong Medical, South Korea). The EUSRA system uses a monopolar 19-gauge needle electrode with an internal active tip, introduced through a curvilinear echoendoscope under real-time ultrasound guidance. Power output is adjustable from 10 to 50 W and is typically applied for 90–300 s depending on lesion size and location. Electrode positioning varies by operator: some center the electrode within the mass, others place the echogenic tip at the distal margin and ablate in sections, and still others ablate sequentially from one side of the lesion to the other. Energy delivery is stopped once tissue impedance rises to roughly 500 ohms or more, a surrogate for adequate coagulative necrosis, and Doppler is used throughout to identify and avoid adjacent vessels.

Habib EUS-RFA catheter (through-the-needle device). An alternative to a dedicated RFA needle is a thin monopolar RF probe (1 Fr, 220 cm) that is passed through the working channel of a standard 19- or 22-gauge FNA needle after tissue sampling, allowing ablation and biopsy in the same session without exchanging instruments. In an early feasibility series of eight patients with pancreatic head lesions (cystic neoplasms and neuroendocrine tumors), this approach was technically feasible and well tolerated. The device is no longer commercially available, but it illustrates the “hybrid” concept of combining tissue acquisition with ablation, and similar combined needles remain an area of device development; outcomes data are still limited to small, single-arm series, and this should be kept in mind when interpreting results.

Device specifications and energy parameters for all systems are summarized in Table 1; safety parameters are summarized separately in Table 2.

Table 1.

Endoscopic Radiofrequency Ablation Devices and Techniques

Modality	Device	Specifications	Technique	Indications	Outcomes	Advantages	Limitations
ERCP-RFA	Habib EndoHPB	7 Fr; 1.8 m; bipolar; 7 W above / 10 W below bifurcation; 90–120 s	Guidewire and fluoroscopy; sequential ablation	Biliary strictures; cholangiocarcinoma	Improved stent patency; survival approximately 10–11 months	Widely available	Thermal injury risk
ERCP-RFA	ELRA	7 Fr; expandable bipolar electrode; approximately 80 °C; 7–10 W	Temperature-controlled ablation under fluoroscopy	Malignant biliary obstruction	Improved survival; stent patency approximately 9–10 months	Temperature feedback improves safety	Hilar anatomy increases risk
EUS-RFA	19G monopolar needle electrode	19G; 10–50 W; 90–300 s	EUS-guided with Doppler; stepwise ablation	PDAC; pancreatic neuroendocrine tumors; cystic lesions/IPMN	Technical success generally above 90%	High anatomic precision	Operator-dependent
EUS-RFA	Hybrid RFA-FNA needle	Combined ablation and aspiration channel	Biopsy followed by ablation in one session	Pancreatic solid tumors (investigational)	High early technical success in small series	Diagnostic and therapeutic in one pass	Limited data; single-study evidence

Table 2.

Technical Considerations and Safety

Parameter	Details
Energy	ERCP: 7–10 W; EUS: 10–50 W
Duration	ERCP: 90–120 s; EUS: up to 300 s
Temperature	Approximately 80 °C (ELRA, temperature-controlled mode)
Guidance	Fluoroscopy (ERCP); real-time ultrasound with Doppler (EUS)
Safety margin	Approximately 2–5 mm from the pancreatic duct and adjacent vascular structures
Antibiotic prophylaxis	Used in most reported series
Common adverse events	Abdominal pain; pancreatitis; cholangitis
Serious complications	Perforation; bleeding (uncommon with appropriate technique)

Abbreviations: ERCP Endoscopic retrograde cholangiopancreatography, EUS Endoscopic ultrasound, RFA Radiofrequency ablation, PDAC Pancreatic ductal adenocarcinoma FNA Fine-needle aspiration

Clinical Evidence by Indication

Malignant Biliary Obstruction

Most of the published experience with ERCP-guided RFA involves patients with unresectable cholangiocarcinoma or pancreatic cancer causing biliary obstruction, typically treated alongside stent placement. In a multicenter registry of 69 patients with unresectable biliary neoplasms, Sharaiha et al. reported a significant improvement in stricture diameter after RFA (p < 0.0001) and a median survival of 11.5 months [8]. A nationwide retrospective study of 84 consecutive RFA applications with the Habib catheter reported a median survival of 10.6 months from the first session, with stent patency of 218 days for metallic stents versus 115 days for plastic stents [9].

A 2025 meta-analysis of nine randomized trials (750 patients total) comparing RFA plus stenting to stenting alone found no difference in three-month stent patency (RR 1.01, 95% CI 0.92–1.11) but improved six-month survival with RFA (RR 0.84, 95% CI 0.73–0.96); post-procedural cholecystitis was more frequent with RFA (5.1% versus 0.3%) [10]. A randomized study using the ELRA catheter before stent placement reported longer stent patency with pre-stent ablation (9.6 months versus 4.5 months, p = 0.029), although overall survival did not differ significantly between groups [11]. A case-control study comparing self-expandable metal stents (SEMS) with and without RFA in unresectable pancreatic cancer found longer median survival with the combination (226 versus 123.5 days, p = 0.010) [12], and a separate retrospective analysis reported that RFA significantly prolonged both survival and stent patency in unresectable extrahepatic cholangiocarcinoma [13]. In a study of in-stent RFA followed by uncovered SEMS placement for recurrent biliary obstruction, time to recurrent obstruction was longer with RFA (117 versus 82.5 days, p = 0.029), though overall survival was unchanged; roughly a third of treatments were interrupted by contact with the existing stent, but adequate energy delivery was ultimately achieved in most patients after repeat application within the same session [14].

Ampullary Lesions

Hu et al. reported a pilot experience of ERCP-guided RFA as primary therapy for inoperable ampullary carcinoma in 23 patients treated between 2012 and 2017, who underwent a median of two sessions (range 1–6) roughly eight weeks apart. Of 18 patients evaluated after treatment, nine had complete lesion resolution and the remaining nine had tumor size reduction exceeding 50%. The six-month re-intervention rate was 36.8%, mean survival was 1,081 days, and six of the twelve surviving patients no longer required biliary stenting. The authors concluded that ERCP-guided RFA was a safe and effective option for tumor debulking in this population [15].

Hilar and Distal Cholangiocarcinoma

Hilar lesions are technically more difficult to treat because of thin ductal walls and adjacent vasculature, and temperature-controlled catheters such as ELRA were developed in part to reduce this risk. In a propensity-matched analysis, temperature-controlled endobiliary RFA was associated with a lower rate of perforation and bleeding compared with conventional power-controlled ablation in inoperable hilar cholangiocarcinoma, without a difference in technical success [16]. Small prospective experiences with the expandable bipolar catheter in hilar and distal cholangiocarcinoma have generally reported three-month stent patency in the 70–80% range with low complication rates, though sample sizes remain small and confirmatory multicenter data are limited.

Pancreatic Solid Tumors, Including PDAC

Armellini et al. described a prospective observational experience using EUS-RFA in patients with locally advanced, unresectable PDAC and found the technique feasible, with acceptable local control and a tolerable adverse-event profile when used as a palliative option [17]. The Habib through-the-needle catheter, delivered via a standard FNA needle, has similarly been used to sample and ablate solid pancreatic lesions in the same session in a small early feasibility series [18]. Choi et al. reported a prospective pilot experience of EUS-RFA in ten patients with small (median 20 mm), benign solid pancreatic tumors, achieving a 100% technical success rate, complete radiologic response in 70% of patients, and a reduction in median tumor diameter from 20 mm to 6.5 mm by the end of follow-up, with associated changes in lesion vascularity and central necrosis on imaging [19].

Pancreatic Neuroendocrine Tumors

EUS-RFA has the most mature evidence base in pancreatic neuroendocrine tumors (PNETs), particularly insulinoma. Oleinikov et al. reported a retrospective two-center experience of 18 patients (7 with insulinoma, 11 with nonfunctional PNETs) and 27 treated lesions, with technical success in 26 of 27 lesions, immediate glucose normalization in all insulinoma patients, and no major complications over a mean follow-up of 8.7 months; mild, transient pancreatitis occurred in two patients [20]. A subsequent multicenter, propensity-matched comparison of EUS-RFA versus surgical resection for sporadic insulinoma found a substantially lower adverse-event rate with ablation (18.0% versus 61.8%) and no severe complications, with clinical efficacy that was slightly lower than surgery (95.5% versus 100%) but recurrences that were generally manageable with repeat ablation; hospital stay was also shorter with EUS-RFA [21]. A 2023 systematic review and meta-analysis pooling 19 studies (183 patients, 196 lesions) reported clinical efficacy of 95.1% for functional PNETs and 93.4% for nonfunctional PNETs, with corresponding overall adverse-event rates of 17.8% and 24.6%; the authors concluded that EUS-RFA has a good safety and efficacy profile for both tumor types, while noting that most included studies were small and retrospective [22]. Across these reports, a safety margin of approximately 2–5 mm from the pancreatic duct and major vessels, confirmed with Doppler imaging, is consistently recommended to limit thermal injury to adjacent structures.

Pancreatic Cystic Lesions and IPMN

Pancreatic cystic lesions, particularly branch-duct IPMN (BD-IPMN), are increasingly discussed as a target for EUS-RFA, but the supporting literature is smaller and less mature than for solid PNETs and is often underrepresented in reviews of this kind despite being listed as a core indication. The rationale is organ preservation: patients with worrisome features or high-risk stigmata under the Fukuoka/Kyoto consensus criteria who decline surgery, or who are poor surgical candidates because of age or comorbidity, may be candidates for ablation rather than resection.

Technique differs from solid-lesion ablation in one important respect: the cyst fluid is aspirated to near-dryness before energy is applied, because intact fluid disperses current away from the cyst wall and reduces ablation efficacy; some protocols then re-instill a smaller volume of saline or contrast to maintain a working space. Impedance and time thresholds used for endpoint determination are correspondingly different from those used in solid parenchymal lesions, and most published protocols favor a single, lower-energy application rather than the stepwise, multi-pass technique typically used for solid tumors.

Barthet et al. reported the first prospective multicenter experience of EUS-RFA in pancreatic neuroendocrine tumors and pancreatic cystic neoplasms, including mucinous cystic lesions and IPMN, with acceptable technical success and a manageable adverse-event profile [23]; long-term follow-up of this cohort, at a mean of approximately 43 months, showed durable reduction in cyst size in most treated lesions, supporting the durability of response beyond the immediate post-procedural period [24]. A subsequent prospective single-arm US trial focused specifically on large (4 cm or greater) BD-IPMN, treating 30 lesions in 25 patients; roughly 79% of lesions achieved at least a 50% reduction in cyst volume and 39% had a complete radiographic response, and next-generation sequencing of fluid resampled at follow-up showed loss of previously detectable KRAS or GNAS mutations in the majority of evaluable lesions, a finding that, while not a validated surrogate for cure, is consistent with a meaningful biological effect of ablation on the neoplastic epithelium [25].

These data are encouraging but still limited to single-arm or non-randomized cohorts with relatively short follow-up relative to the natural history of IPMN, which can span many years. Patient selection, the optimal number and spacing of treatment sessions, and the durability of mutation clearance all remain open questions, and EUS-RFA for cystic lesions should currently be considered an option for carefully selected patients who are not candidates for, or who decline, surgical resection, rather than a routine alternative to either surveillance or resection.

Safety and Adverse Events

Adverse events with both ERCP- and EUS-guided RFA range from mild, self-limited abdominal discomfort to more serious complications, including infection, acute pancreatitis, cholangitis, perforation, and bleeding. Across the series reviewed above, careful attention to power and duration settings (summarized in Table 2) appears to reduce the frequency of these events, and most authors administered prophylactic broad-spectrum antibiotics before the procedure [20, 21, 25]. Maintaining a safety margin of approximately 2–5 mm from the main pancreatic duct and adjacent vascular structures, confirmed with Doppler imaging before activation, is the most consistently cited technical measure for reducing the risk of thermal injury to surrounding tissue. Table 1 summarizes device specifications and outcomes by indication, and Table 2 summarizes the energy and safety parameters discussed throughout this review.

Conclusion

Endoscopic RFA, delivered via either the ERCP or EUS route, is a reasonable adjunct in the management of unresectable pancreaticobiliary malignancy and is being explored increasingly for pancreatic cystic lesions and IPMN. Technical success is generally high and most adverse events are mild, but the underlying evidence base is dominated by small, single-arm, or retrospective studies with heterogeneous energy protocols and limited long-term follow-up, particularly for cystic lesions, hilar cholangiocarcinoma, and hybrid sampling-and-ablation needles. Prospective, ideally randomized and multicenter, studies with standardized technique and longer follow-up are needed to define patient selection criteria and confirm durable oncologic benefit before these indications can be considered routine.

Authors’ contributions

N.T. conceived the study and designed the review. A.B. conducted the literature search, drafted and contributed to revision of the manuscript. J.L. assisted with manuscript editing and formatting. All authors reviewed, revised, and approved the final manuscript.

Funding

No funding was received for this work.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.