Radiofrequency ablation (RFA) has been shown to induce antigen-presenting cell infiltration and enhance systemic antitumor T-cell immune response as well as tumor regression in hepatocellular carcinoma (36). The tumor necrosis generated could be an antigen source for the immune system and it has been demonstrated that RFA determines a weak but detectable immune response which involves the activation of macrophages and the release of inflammatory cytokines (34). An early increase of cytokine IL6, followed by a delayed elevation of the serum levels of chemokines CXCL11, CXCL5, and CXCL1 was recently demontrasted in patients with pancreatic cancer and cholangiocarcinoma undergoing ERFA (37). However, the systemic immune response detected after ERFA was not specifically related to the endobiliary ablation and it was attributed to a general inflammatory response (37). Most likely, the immunogenic effects of RFA occur at the tumor site where the necrotic neoplastic tissue induces severe inflammation which can determine immune-mediated tumor destruction by neutrophils, macrophages, dendritic cells, natural killer cells together with B and T lymphocytes (33, 37). It has been speculated that the immune-mediated tumor destruction is not triggered by necrotic neoplastic tissue, but it is induced by the immunostimulatory and inflammatory factors present in the sub-vital tissue surrounding the ablated necrotic area (33). This could be particulary true for biliary strictures treated with ERFA where there is no certainty of complete tumor destruction since the width, the depth and the length of the ablation are not foreseeable, as demonstated experimentally (38–43). The local immunogenic response could potentially enhance tumor suppression and be responsible for the improved survival of patients with MBO and unresectable CCA who undergo ERFA (20–23) ( Figure 1 ).

To date, there are two ERFA catheters which have been approved for clinical use ( Table 2 ). The HABIB catheter (Boston Scientific, Marlborough, MA, USA) is a power-controlled 8-French bipolar probe, 180 cm long, with two circumferential electrodes placed 8-mm apart on the distal tip of the catheter to achieve an ablation length and depth of 20-25 mm and 3-5 mm, respectively (17, 43, 44). The HABIB catheter can be connected to different RFA generators, among which the most frequently used are the ERBE electrosurgical generators (Erbe,Tübingen, Germany). The recommended settings are effect 8 and 10 Watts (W) for ablation in the common bile duct and 8W for ablating strictures at the biliary bifurcation, near the cystic duct and the ampulla (38, 39). The ELRA catheter (STARmed, Goyang, Korea) is a 7-French temperature-controlled bipolar probe, 175 cm long, with two to four circumferential electrodes in its distal tip which are placed at different lengths. There are four distinct types of ELRA probes which accomplish diverse coagulation lengths (11, 18, 22, and 33 mm) along with circumferential ablation depths between 6 and 8 mm and median ablation depth of 4.0 mm (39, 40, 43, 45). The ELRA catheter operates only with the VIVA comboTM RFA generator (Taewoong Medical, South Korea), which allows presetting the target temperature and automatically stopping the procedure if this is exceeded during the ablation time (45). The recommended settings are a target temperature of 80°C and a power of 7 W or 10 W, on the basis of the type of catheter used as well as the location of the biliary stricture (39, 40, 45). The ELRA catheter differs from the Habib probe due to its different length and its temperature sensor on the tip of the device which provides a temperature-controlled ablation. Theoretically, these features offer the advantage to properly treat biliary strictures of different lengths and to reduce the risks of injuring the biliary ducts (45, 46).

ERFA can be performed at the time of ERCP or PTC before biliary drainage in patients with MBO and strictures due to different neoplastic etiologies (11, 47, 48). Both approaches require cholangiography to properly visualize and measure both the length and the caliber of the stricture, before placing the wire-guided ERFA catheter inside it, under fluoroscopic monitoring (41, 45). The tip of the probe with the electrodes has to be positioned in direct contact with the target tissue. This is crucial for tissue destruction using either one of the devices, with a linear relationship between depth of ablation, preset power and established time of ERFA delivery (40, 49). Usually, each ablation lasts 60-120 seconds, with an average time of 90 seconds (38–41). In the case of long strictures (> 15 mm) the ablation needs to be repeated, without overlapping the treated segments when using the HABIB catheter; on the other hand a different length of the ELRA catheter can be choosen, avoiding repeated ablations for strictures up to 30 mm in length (41, 45). However, two or more ablations are always required when there is a complex hilar stricture, which requires separate treatment of both the right and left hepatic ducts (47). When ERFA is repeated, a 60 s resting period is recommended between applications. After removing the ERFA catheter, the bile duct is swept by using a retrieval balloon to remove residual necrotic tissue and a cholangiogram is obtained to rule out possible complications (24). Eventually, a biliary stent is placed to ensure long term biliary drainage, because of the stricture and the possibility of additional segmental biliary stenosis that ERFA can cause (40). Plastic stents are indicated if periodic ERFA sessions are planned at regular intervals, while metal stents are placed only when a single ERFA is forecasted and performed at the time of biliary drainage (50).

The therapeutic efficacy of ERFA could be affected by the morphology of the biliary stricture and its location (11, 47). Since the electrodes of the ERFA catheter need to be in strict contact with the target tissue, some features of the biliary stricture can affect the results of the ablation. The narrower the stricture, the higher the amount of energy delivered inside the stenosis: a powerful ablation not only causes deep and irregular coagulative necrosis, but also results in ineffective tissue ablation and increased risk of injury to the duct (18, 51). Similarly, when the biliary stricture is short, irregular or mushy and loose, ERFA could be unsuccessful because of uneven contact between the electrodes of the probe and the target tissue (22, 47, 52). Furthermore, the electrodes of the ERFA catheter can overlap the stricture, and ablate the normal bile duct beyond the biliary stenosis. This usually happens during the ablation of short strictures, but it can also occur with long stenosis when overlapping consecutive ablations are performed (22, 47, 52). In both cases ablation of the normal bile duct develops scar tissue, which expands the length of the original stricture and increases the risk of stent occlusion (40).

Anatomic characteristics of the intrahepatic and hilar biliary ducts might affect the efficacy of ERFA and/or increase the risk of complications. Bile ducts angulation in the hilum can impair ERFA treatment because straight and rigid endobliary catheters may not pass the angulations and/or the tip of the probe may not mantain the required tight contact with the target tissue (53). Hilar and intrahepatic biliary ducts have a subtle wall that is more susceptible to thermal damage, which can extend to surrounding structures, even in the presence of a tumor mass (42, 54). Strictures located in the hilum are close to both portal and hepatic veins as well as hepatic arteries: the blood flow acts as a cooling circuitry (heat-sink effect), which may prevent the ERFA catheter to deliver the proper energy needed to obtain effective tissue ablation (11, 49, 53). Intrahepatic strictures may be difficult to ablate completely because the ERFA catheter cannot pass easily them or the stenoses are too numerous to be all treated effectively (53). In these cases, selective ablation of dominant strictures is performed because complete treatment is not feasible (54).

Several possible solutions to the above mentioned technical issues have been proposed. A preoperative road map with abdominal MRI (magnetic resonance imaging) and MRCP (magnetic resonance cholangiopancreatography) is recommended to accurately assess the tumor surroundings and evaluate the relationship of the target biliary stricture with the surrounding vascular and biliary structures, especially for the treatment of hilar and intrahepatic stenosis (50). Before ERFA, both the length of the stricture and the thickness of the biliary wall should be measured by using intraductal or endoscopic ultrasonography, especially if there is no apparent tumor mass on preoperative imaging (18, 55, 56). This information maximizes the efficacy of ERFA, reducing the risk of thermal injury by means of proper settings of the RFA generator and the duration of ablation, respectively (38–42, 45, 49). Patient-tailored settings may achieve better clinical outcomes for ERFA, which can be ultimately adapted to the native anatomy and the tumor mass (39). The temperature reached by the target tissue during ERFA correlates with the thermal damage of the bile ducts. Therefore, the novel temperature-controlled ERFA system could avoid unintended thermal injury of the biliary wall and the surrounding structures (45, 46). ERFA usually lasts 90-120 seconds. During this time the position of the electrodes may change inside the stricture provoking unintended thermal injury to the normal biliary wall. Therefore, it has been proposed to perform two consecutive 60 seconds ablations with an interval time of 60 seconds which is useful for checking the position of the electrodes by means of fluoroscopy and detecting the possible onset of adverse events, such as bleeding (53). The best way to correctly place the electrodes of the ERFA catheter inside the biliary stricture is to position the probe after direct visualization and evaluation of the stenosis using the peroral digital cholangioscope (57–59). Subsequently, another cholangioscopy evaluates the efficacy of the ablation and rules out possible immediate complications, such as bleeding and perforation (60). Placement of metal stents has been recommended to prevent bile duct injury, because they assure an immediate decompression of the biliary tree and a cooling effect on the ablated tissue by means of a copious biliary flow (23). Placement of fully covered SEMS has been suggested to avoid septic consequences of inadvertent bile duct injury (61, 62).A different technique can be considered for the local treatment of the biliary strictures if the risk of collateral damage induced by ERFA is classified as too high, at the time of preoperative road map (18).

The major advantages of ERFA are simplicity and low cost, without many major adverse events and very few controindications. The latter include the presence of cardiac devices, coagulation disorders, and ascites, as well as pregnancy (11, 63). There is considerable variability in the reported incidence of the adverse events after ERFA that can range from 7% to 48%, averaging the data of four previous published reviews (42, 56, 63, 64). This variability can be due to the fact that some adverse events after ERFA are not strictly related to it, but are the possible complications after ERCP or PTC, and therefore they might not have been reported (47). Other explanations reside in different etiology, location and morphology of the stricture; degree of thickness of the bile duct wall; contiguity of vascular structures with the biliary stenosis; variance in energy settings and duration of ablation; type of biliary stents (plastic or metal) placed after ERFA (64). The majority of patients complain of abdominal pain, which occurs in almost 50% of cases and it is self limited (10, 14, 65–67). Reported pancreatobiliary adverse events are pancreatitis, cholangitis, cholecystitis and minor bleeding (10, 14–16, 18–20, 22, 23, 44, 46, 50, 55, 61, 65, 66, 68–74). These are the typical adverse events that can occur after ERCP or PTC (5–8, 75). However, a higher number of cholecystitis and cholangitis have been reported, especially in patients with hilar strictures treated with ERFA (10, 22, 23, 50, 65, 68, 69, 73, 76). The incidence of cholecystitis requiring percutaneous drainage after ERFA has been estimated to be between 2%–4% and it is significantly higher than that reported after standard biliary drianage; similarly, cholangitis seems to occur more frequently after ERFA and subsequent placement of biliary stents (2-8%) (24). A possible explanation for the onset of suppurative cholecystitis could be the obstruction of the cystic duct, as a consequence of its thermal injury due to edema or tissue destruction when ERFA is delivered too close to the opening in the bile duct (10, 65). However, cholecysistis is not always reported after ERFA (18), and it has been hypothesized that the type (plastic vs metal) and the number (1-2 vs multiple) of biliary stents could be associated with a higher risk of cystic duct blockage (13). Since the reported total number of cases of cholecystis remains low and this complication resolves in few days with percutaneous gallbladder drainage and/or antibiotics, ERFA with subsequent biliary drainage is considered safe, even when the biliary stricture is close to the opening of the cystic duct (65). The high frequency of cholangitis has been initially attributed to necrotic debris which can remain in the bile ducts after ERFA with early subsequent obstruction of biliary stents; to avoid this possible complication the bile ducts are swept with an extraction ballon after ERFA and before stent placement (24, 48, 65, 77). Another possible cause of cholangitis is the ablation of the normal bile duct beyond the stricture, which sometimes occur because of technical and/or anatomical difficulties determining the onset of iatrogenic strictures of the bile ducts which could not be properly stented (40, 47). To reduce the risk of unintended strictures an accurate measurament of the stricture is required, especially if the biliary stenosis is long and requires ovelapping ERFA (18, 55, 56, 78). Few life-threathening adverse events have been reported (20, 36, 50, 62, 79–81). Therefore, it is important for the biliary endoscopist to be aware of these complications. Seven cases of biliary perforations occurred after endoluminal ablation of narrow biliary strictures, two of which were dilated before performing two overlapping ablations (20, 62, 73). After ERFA late severe melena developed in two patients and this was due to the rupture of a pseudoaneurysm originating from an artery, which was too close to the electrodes of the ablation catheter (79, 81). Six cases of delayed hemobilia were reported 4-6 weeks after ERFA and two of them were fatal because of hemorrhagic shock (41, 50). Liver infarction due to arterial thrombosis was diagnosed in a patient 3 days after ERFA of a stricture of the right hepatic duct: this complication was attributed to the proximity of the biliary stricture with a branch of the right hepatic artery (50). Vascular as well as biliary injuries ending in severe complications are mostly related to severe thermal injury of the bile duct inside and beyond the stricture which extends to the surrounding vascular and biliary stricture (50, 79, 81). Furthermore, aberrant angiogenesis after ERFA could explain delayed spontaneous hemobilia (41, 50). Placement of a SEMS after ERFA could be an effective method for preventing the onset of late bleeding and biliary fistulas. It has been hypothesized that the high radial force of SEMS may have both a tamponade and hemostatic effect on the oozing from the necrotic tissue resulting after ERFA (25, 78). The rapid flow of bile through the strictures ensured by SEMS could have a cooling effect preventing deep bile duct injuries (23). Hyperkalemia was the cause of a sudden non-fatal cardiac failure in a patient with chronic kidney disease who underwent ERFA for the treatment of a biliary stricture at the time of biliary drainage (80). Another three cases of non-lethal heart failure occurred in two patients with a history of coronary heart disease and hypertension within 24 hours after ERFA (20, 74). Finally, a case of fatal hepatic coma, a left bundle branch block, and a few cases of liver abscess have been reported, especially after ERFA at the time of PTC (16, 19, 21, 22, 50, 54).

Strict patient-selection and ablation with customized settings (according to the location of the biliary stricture and the comorbidity of the patient) have been proposed to reduce the incidence of severe complications (20). Careful postoperative follow-up is necessary, and evaluation of the results of ERFA with cholangioscopy has been recommended (47, 60, 66).

Maintaining the patency of biliary stents guarantees the administration of chemotherapy without interruption. Despite the use of biliary SEMS, recurrent jaundice and cholangitis tend to occur because of reobstruction of the biliary stents due to tumor growth (7, 8). Several studies have demonstrated the beneficial effects of ERFA on stent patency, which is usually prolonged by approximately 2 months (20–23). The effects of ERFA on stent patency has been investigated by the majority of the cited comparative studies (15, 18, 23, 53, 54, 67, 68, 70). It seems that ERFA has the capability to prolong the patency of uncovered metal stents inducing a reduction in the tumor mass, which is associated with slowed endobiliary neoplastic growth and improved bile flow (52, 54). The decreased risk of sludge and/or biofilm formation could also explain the prolonged patency of plastic stents fter ERFA (54). As mentioned above, three single arm studies confirmed that ERFA can be performed safely at the time of biliary drainage either by means of ERCP or PTC and followed by placement of plastic or metal stents (44, 61, 74). The advantage of using plastic stents is that they permit repeated sessions of ERFA at scheduled times and this protocol seems to be beneficial for patients with unresectable CCA (15, 18, 44, 68). The impact of ERFA on biliary stent patency has been confirmed by a recent meta-analysis whose data demonstrated the superiority of ERFA plus stenting over stenting alone, independent of the stent type used (plastic vs metal) (10). However, these data are still controversial since it has been reported that ERFA has no effect on prolonging patency of both metal and plastic stents, respectively by a retrospective study and two randomized controlled trials (53, 54, 69). Similar doubtful and inconclusive results were obtained by a recent meta-analysis whose authors were unable to perform a pooled analysis of avalaible data and just reported that only three of five studies evaluated showed a beneficial impact of ERFA on stent patency (65). Biliary plastic stents need to be exchanged and this can be performed respectively, on schedule every three months or on demand (i.e. at the occurrence of signs and/or symptoms of obstruction) after the second session of ERFA scheduled at the time of first 3-month endoscopic follow-up (18, 68). When plastic stents are exchanged, a repeat ERFA session can be performed. The need of reintervention could be decided on the basis of the results of cholangiography and/or intraductal ultrasonography (IDUS), which can measure the caliber and the width of the bile duct (18, 48, 56, 66). ERFA should be repeated when IDUS detects a significant increase in tumor thickness and a reduction in the bile duct diameter at the site of the previously treated stricture (18). Another technique used to monitor the results of ERFA is cholangioscopy which can also guide the correct placement of the ERFA catheter inside the targeted biliary stricure (58–60). Ideally, plastic stents are indicated when multiple sessions of ERFA are scheduled in patients with a locally advanced CCA without metastases and in good functional status (24). On the other hand, SEMS are recommended when only a single session of ERFA is planned and their use has been advocated to reduce the risk of late bleeding and biliary fistulas (23, 25, 78). Moreover, SEMS are the preferred stents after the execution of ERFA at the time of PTC, which is usually performed to treat intrahepatic unresctable CCA (23, 48, 65, 67, 70).

The most valuable effect of ERFA is its impact on the overall survival of patients with unresectable CCA undergoing biliary drainage and stent placement. The above mentioned comparative studies as well as the cited meta-analysis investigated the impact of ERFA on overall survival and improved functional status of patients (15, 18, 23, 53, 54, 65, 68, 70). The overall survival of patients with unresectable CCA is significantly improved after ERFA plus stenting, with a pooled mean survival of 374 days vs 263 days of those treated only with stent placement at the time of biliary drainage (15, 18, 23, 68, 70). Similar data were obtained by the meta-analysis which reported a median survival of 294 days in patients undergoing ERFA vs 216 days in those who received only a biliary stent, independent from the type of stent placed (65). As already mentioned, the improved survival of patients undergoing ERFA could be due to the local immunogenic response, which potentially enhances tumor suppression and decreases the tumor burden delaying neoplastic progression inside as well as outside the bile duct (20–23). The direct action on the tumor and the induced local and systemic immune mechanisms could explain the favorable impact of ERFA on overall survival of patients undergoing endobliary ablation (65, 68). Only two retrospective studies reported no difference in overall survival between patients undergoing ERFA and those treated only with biliary stent placement (53, 54). There are several possible explanations for these controversial results: the anatomy of the biliary ducts, which could have been too angulated for adequate ablation; the cooling effect due to the blood flow of the surroundings vessels which could have prevented sufficient ablation of the tumor; the type of CCA treated, since Bismuth III and IV are characterized by the presence of multiple strictures, which could not be ablated as a whole, invalidating the efficacy of ERFA; the placement of SEMS which could have hidden the beneficial effects of ERFA; the use of different probes and generator settings which could have affected the outcomes, especially in patients with hilar CCA undergoing ERFA (53, 54). Besides overall survival, ERFA seems to have also a beneficial effect on the functional status of the patients undergoing endobiliary ablation. Several studies reported rapid improvement of the jaundice and increased albumin values which translated to a better functional status and higher Karnofsky Performance Scale (KPS) scores in comparison with patients treated only with stent placement (18, 23, 68, 70). ERFA, cancer stage, Bismuth type I-III, level of serum albumin near normal and the administration of adjuvant chemotherapy could be positive prognostic factors that have a beneficial cumulative impact on the overall survival of patients with unresectable CCA (23, 65, 68). Among these, adjuvant chemotherapy has been proven to be the most effective and its combination with ERFA could be the best option to improve the overall survival of patients with advanced CCA (65).

It has been postulated that the thermal cell injury induced by ERFA could increase the cytotoxic effect of chemotherapy, especially in the case of intrahepatic CCA where the endobiliary ablation is often sublethal (53). Moreover, some data suggest that stent patency is shorter after ERFA without chemotherapy, and failure of its administration can be considered a risk factor for stent occlusion in patients with unresectable CCA undergoing endobiliary ablation (23, 53). The possible advantage of the combination between ERFA and adjuvant chemotherapy has been investigated in three studies, two of which were retrospective and the other one was a randomized controlled trial (16, 52, 55). In the two retrospective studies patients undergoing combination therapy were compared with those treated only with chemotherapy after biliary drainage with the aim of evaluating the impact of combination therapy on both the overall survival and the progression free survival of patients with unresectable CCA ( Table 6 ) (16, 52). The superiority of combination therapy over ERFA alone was then demonstrated by a randomized controlled trial which investigated the effect of the consecutive administration of ERFA and a novel anti-cancer drug in improving both overall survival and progression free survival of patients with locally advanced unresctable CCA (55). All the data presented in these studies support the efficacy of the additional effect of ERFA on chemotherapy, with an average median survival of 16.6 months compared to 10.3 months of patients undergoing only chemotherapy (16, 52, 55). Similarly, median progression free survival (PFS) was improved in patients undergoing chemotherapy after ERFA (16, 52). These advantages were clear for locally advanced CCA, but became less evident in patients with metastatic CCA, for whom the combination of ERFA and chemotherapy did not significantly increase both median survival and PFS in comparison to patients undergoing chemotherapy alone (16, 52). The combined therapy also had a beneficial impact on the functional status and the quality of life of patients with a prolonged high KPS scores after ERFA (55). No major side effects of both treatments and no increase in adverse events were reported with the combination of ERFA together with adjuvant chemotherapy (16, 52, 55). Therefore, the results of these three studies support our change of approach in patients with locally advanced CCA undergoing biliary drainage, especially if they have a life expectancy of at least 6 months: if possible, they should undergo ERFA before starting adjuvant chemotherapy (82).