Percutaneous endobiliary radiofrequency ablation and stent placement for unresectable malignant biliary obstruction: a propensity score matching retrospective study

Abstract

Background

Whether endobiliary radiofrequency ablation (EB-RFA) changes the standard role of stent placement in treating unresectable malignant biliary obstruction (MBO) remains unclear. The aim of this study is to compare percutaneous EB-RFA and metal stent placement (RFA-Stent) with metal stent placement alone (Stent) in treating unresectable MBO using a propensity score matching (PSM) analysis.

Methods

From June 2013 to June 2018, clinical data from 163 patients with malignant biliary obstruction who underwent percutaneous RFA-Stent or stenting alone were retrospectively analyzed using a nearest-neighbor algorithm to one-to-one PSM analysis to compare primary and secondary stent patency (PSP, SSP), overall survival (OS) and complications between the two groups.

Results

Before matching, for whole patients, RFA-Stent resulted in longer median PSP (8.0 vs. 5.1 months, P = 0.003), SSP (9.8 vs. 5.1 months, P < 0.001) and OS (7.0 vs. 4.5 months, P = 0.034) than the Stent group. After matching (54 pairs), RFA-Stent also resulted in better median PSP (8.5 vs. 5.1 months, P < 0.001), SSP (11.0 vs. 6.0 months, P < 0.001), and OS (8.0 vs. 4.0 months, P = 0.007) than Stent. RFA-Stent was comparable with Stent for complication rates. In Cox analysis, RFA-Stent modality and serum total bilirubin level were independent prognostic factors for PSP. RFA-Stent modality, performance status score and combination therapy after stent were independent prognostic factors for OS.

Conclusion

Percutaneous RFA-Stent was superior to Stent in terms of PSP, SSP, and OS in selected patients with unresectable MBO.

Background

Malignant biliary obstruction (MBO) comprise a group of tumors, particularly cholangiocarcinoma and pancreatic carcinoma, and resection is considered to be the only curative modality [1]. However, many patients with MBO are at advanced stages at the time of presentation and have lost surgical opportunities. The primary therapeutic goal for these patients is to relieve biliary obstruction and clinical symptoms [2]. Compared with plastic stents, self-expandable metal stents (SEMS) are considered as standard palliative care for MBO due to their precise positioning, high technical and clinical success rates [3]. However, biliary stent placement can only relieve mechanical obstruction with limited effects on controlling local tumor growth, and were prone to re-clogging, so most patients died within six months [4].

Radiofrequency ablation (RFA) is considered to be a radical, technically mature and minimally invasive treatment for a variety of solid tumors [4]. However, cholangiocarcinoma and other tumors growing along luminal structures are not suitable for current traditional percutaneous ablation techniques due to the difficulty of visualizing and targeting the tumor through percutaneous approaches. Intraluminal radiofrequency ablation with an endobiliary catheter has been applied for endobiliary radiofrequency ablation (EB-RFA) within the recent decade [5, 6]. The catheter in EB-RFA procedure is a bipolar catheter that can be guided by a guide wire along the bile duct tree across malignant strictures so that ablation energy can be applied to the stenosis prior to stent placement [6]. The ablation energy is transferred directly to the surrounding tissue, causing coagulation necrosis [7]. We have previously reported that percutaneous EB-RFA is effective and safe for treating MBO caused by different etiologies [8,9,10]. There were many studies on percutaneous EB-RFA plus stenting (RFA-Stent) versus stenting alone (Stent) for MBO. Most studies showed percutaneous EB-RFA has a potential to results in stent patency and survival benefits [11,12,13]. In contrast, in our previous retrospective study and a recent small size randomized study, the results showed RFA-stent improved stent patency compared to stent placement alone, but had no effect on survival [14, 15]. However, these studies were limited by relatively small study populations and have yielded conflicting results. Also, the therapeutic outcomes were compared without correction for potential confounding factors in the retrospective studies that might affect the outcomes of RFA-Stent and Stent. Recently, a prospective randomized study suggested that RFA-Stent using an endoscopic approach was superior to Stent in the treatment of unresectable MBO [4]. While, no multicenter prospective randomized tries involving more patients or propensity score matching retrospective studies on percutaneous RFA-stent versus stent for unresectable MBO have been reported to confirmed the results. In this one-to-ne propensity score matching (PSM) retrospective study, we aimed to compare the safety and efficacy of percutaneous RFA-Stent with Stent in treating MBO.

Methods

Study design

Between June 2013 and June 2018, the electronic medical records from consecutive MBO patients who underwent RFA-Stent or Stent were retrospectively reviewed (Fig. 1). Patients who received percutaneous EB-RFA + stent placement were included in the RFA-Stent group, while patients who underwent stent placement only were included in the Stent group. The inclusion criteria were: MBO clinically or histologically confirmed by diagnostic imaging (e.g., computed tomography (CT), abdominal magnetic resonance imaging (MRI)), biopsy, or a previous surgical procedure; and an unresectable tumor or refusal of surgery. age ≥ 18 years; Patients with a performance status (PS) score > 2, those with serious medical comorbidities, such as a severe coagulation disorder, or insufficient renal or serious cardiac function, those with missing data and those with secondary malignancies were excluded (Fig. 1). This study was approved by the Institutional Review Board of our hospital. The requirement for informed consent was waived given its retrospective design.

Fig. 1

Flowchart shows the patient selection criteria. RFA-Stent: percutaneous endobiliary radiofrequency ablation and self-expandable metal stent placement; Stent: stent placement alone

Treatment protocol

In RFA-Stent group, the procedures were performed under local anesthesia and no conscious IV sedation. A 22G needle (COOK) was advanced through the abdominal wall and liver into a bile duct under ultrasound or fluoroscopic guidance. Contrast medium was then injected to fill the bile ducts, demonstrating the degree of dilation, biliary obstruction site, its length and diameter. If one side percutaneous transhepatic cholangiography (PTC) indicated that the contralateral bile duct was not connected, bilateral bile duct drainage would be required. Then, an 8 French bipolar RF catheter (Habib™ PERF; EMcision Ltd, London, UK) was placed at the malignant stenosis site with the guide wire. The RF catheter was connected to a standard high-frequency generator (RITA 1500X RF generator, AngioDynamics, Manchester), with 10 W applied for 90 s at one section (Fig. 2). EB-RFA was performed section by section as follows: for a bile duct obstruction no longer than 4 cm, two EB-RFAs were performed along the segmental obstruction without overlapping the already treated segments, which can theoretically produce a necrosis area with a maximum length of 3.9–4.6 cm; To avoid a further increase in tissue temperature that would probably result in charring and bile leak, a 60-s interval was allowed between ablations. For a bile duct obstruction longer than 4 cm, three EB-RFAs were performed along the segmental obstruction without overlapping the already treated segments and a 60-s interval after each ablation, which can theoretically produce a necrosis area with a maximum length of 5.9–6.9 cm; For bilateral high level obstruction, both intrahepatic obstructed bile ducts required EB-RFA. The main aim of EB-RFA is to improve luminal stenosis and bile duct obstruction rather than completely ablate tumors. Cholangiography was performed immediately to determine the technical success of EB-RFA and whether bile duct recanalization had been achieved. Next, an uncovered SEMS (Wallstent; Boston Scientific, Boston) of suitable length was placed. Generally, the diameter and length (40–60-80 mm) of SEMS were selected according to the individual radiologist’s preference and the manufacturer’s protocol. A stent with a diameter of 8 or 10 mm is used for patients with unilateral obstruction and a stent with a diameter of 6 mmm is used for patients with bilateral obstruction requiring the implantation of two stents via the "stent-by-stent" technique. Repeat PTC was performed immediately to determine the location of stent placement. After the procedure, a pig-tail catheter was implanted for easy management of the next PTC and complications such as bleeding. Three to four days after stent placement, PTC was performed to re-confirm biliary stent patency and remove the drainage catheter.

Fig. 2

A 61-year-old patient presence history of unresectable pancreatic cancer was treated with stent placement receiving percutaneous RFA-Stent for MBO. Axial CT images in axial and coronal CT images (A-C) shows the dilated bile ducts. D Fluorescence image shows percutaneous transhepatic cholangiography. E A guide wire was then passed through the stenosis. F-G An 8 Fr. (2.6-mm), 1.8-m bipolar RF catheter was then advanced over the wire with the tip of the catheter placed across the malignant stricture. Percutaneous radiofrequency ablation for MBO. H Stent in position. I A cholangiogram showing stent patency four days after the RFA treatment. AC: ablation catheter tip

In Stent group, only stent placement was performed. The choice of stent diameter and length, and stent placement and drainage technique were exactly the same as in the RFA-Stent group.

Assessment and follow-up

The primary endpoint was primary stent patency (PSP; defined as time from the initial stent placement to recurrence biliary obstruction (RBO)). RBO was defined as a composite endpoint of either migration or occlusion resulting from tumor ingrowth /overgrowth, new stricture, sludge or unknown [14]. Secondary endpoints were overall survival (OS; defined as time from the initial stent insertion to death from any reasons or last follow-up.), secondary stent patency (SSP; defined as the cumulative duration of stent patency, which corresponded to the absence of recurrent symptomatic biliary obstruction.), technical success, complications, and length of hospital stay.

Technical success of EB-RFA was defined as significant improvement of biliary obstruction after ablation, which was determined by two criteria, i.e., injection of contrast into the strictures flowing fluently after ablation and successful stenting without a balloon dilation assistant. Technical success of stenting was defined as passage of the stent across the stricture with good radiographic positioning and flow of contrast and/or bile through the stent. The technical success of EB-RFA or stenting was assessed by cholangiography during the procedure. Clinical success was defined as improvement of symptoms, such as jaundice and pruritus, and total bilirubin levels to less than half or less than the normal upper limit within 14 days of the intervention [14]. Complications were classified as major or minor and assessed using the Society of Interventional Radiology classification system.

Medical imaging (contrast-enhanced CT or MRI) was performed to characterize the tumor and stent patency (biliary obstruction) every 6–8 weeks during the first year and every three months thereafter. When recurrent biliary obstruction was suspected on the basis of clinical symptoms or elevation of direct bilirubin levels, contrast CT or MRI was performed to determine whether the intrahepatic or extrahepatic bile ducts were dilated. When stent occlusion (or RBO) was identified during follow-up, the same treatment strategy of either EB-RFA (with or without a new stent) in RFA-Stent group or stent placement in Stent group was recommended as previously described for the first intervention. If RFA-Stent or Stent failed, percutaneous transhepatic cholangiography and drainage were performed. In addition, during follow-up, combination therapy for tumors, including systemic therapy (chemotherapy, targeted therapy, immunotherapy), locoregional therapy (chemoembolization, ablation, radioactive seed implantation), and surgery (is necessary because of complications), was recommended by the multidisciplinary tumor team based on the characteristics of the tumor, hepatic function, and general condition of the patient.

Statistical analysis

Adjusting treatment groups by a PSM method was performed to address the selection bias of potential confounders between two groups. Two groups were propensity scored matched using the nearest neighbor method to one-to-one matching (caliper = 0.2) without replacement. Independent variables used for propensity model were sex, age, performance status score, pathology confirmed, etiology, level of biliary obstruction, length of initial stricture, previously therapy for primary tumor, previous cholangitis, combination therapy, total bilirubin (TB) and gamma-glutamyl transpeptidase (GGT). Given that P-values can be biased by population size, the effect size was reported as the standardized mean difference before and after matching: |effect size|< 0.1 indicated very small differences, |effect size|< 0.3 indicated small differences, |effect size|< 0.5 indicated moderate differences, and greater than 0.5 indicated large differences [16].

We use Pearson χ2, Fisher’s exact tests, and continuity correction and independent samples t to compared variables between two groups as appropriate. PSP, SSP and OS were estimated using the Kaplan–Meier method. In univariate and multivariate analyses, potential prognostic factors, including all baseline covariates for PSP, OS and SSP, were analyzed using Cox proportional hazards models for all patients.

All statistical analyses were performed using GraphPad Prism software (version 6.01, GraphPad Software, La Jolla California, USA) and IBM SPSS Statistics software (version 24.0, IBM Corp., Armonk, NY). Differences were considered statistically significant when a P-value was < 0.05.

Results

Patients

A total of 163 eligible patients, 90 (mean age 60.7 [range 33–85] years; 56 men, 34 women) who underwent RFA-Stent and 73 (mean age 63.9 [range 33–91] years; 39 men, 34 women) who underwent Stent for MBO were included. The baseline demographic and clinical data for the two groups are shown in Table 1. RFA-Stent group contained more patients aged ≤ 65 years (P = 0.098), more with cholangiocarcinoma and other diseases (P = 0.097) and with low-level biliary obstruction (P = 0.002); patients in this group had also received less treatment for primary disease (P = 0.083). PSM created 54 matched pairs of patients. The baseline data after matching are shown in Table 2. In this matched cohort, the selection bias attributable to potential confounders was well balanced and there was comparable between the two groups. Histograms and scatterplots of the matching output are displayed in Supplemental Fig. S1.

Table 1 Baseline patient characteristics before propensity score matching

Table 2 Baseline patient characteristics after propensity score matching

Treatment details

In RFA-Stent group, bilateral stents were placed at the first time of procedures in 19 patients (21.1%). A total of 109 biliary strictures were ablated and followed by 109 stentings. The initial success of EB-RFA was 98.2%) per stricture (107/109) and 97.8% per patient (88/90). Two patients underwent stenting with a balloon dilation assistant because there was no significant improvement in biliary obstruction after ablation. The initial technical success of stenting was 97.2% per stricture (106/109) and 96.7% per patient (87/90). One patient underwent repeat stenting because of stent migration, and two underwent balloon dilatation because the stent was not fully expanded and the duct remained partially obstructed. A median of one EB-RFA procedure (range 1–5) was performed during the study period. Ten patients underwent two ablations, with no new stenting in seven patients, and three underwent three ablations without new stenting because of RBO. One patient underwent five ablations with only one new stenting. The clinical success rate was 85.6%.

In Stent group, bilateral stents were performed in nine patients (12.3%). Totally, 82 stenting procedures were performed. The initial success of stent insertion was 98.8% per stricture (81/82) and 98.6% per patient (72/73). One patient underwent balloon dilatation because the stent was not fully expanded and the duct remained partially obstructed. During the study period, a median of one stenting procedure (range 1–2) was performed, and five patients (6.8%) underwent repeated stenting because of RBO. The clinical success rate was 84.9% (62/73). Technical and clinical success rates of stent insertion between the two groups were comparable (P = 0.767 and P = 0.911, respectively).

Complications, hospital stay, and causes of RBO

The RFA-Stent group showed slightly higher complication rate than did Stent group (Table 3). In RFA-Stent group, two patients died of cholangitis and septic shock within 30 days of the procedure. In Stent group, four patients died; one from cholangitis and septic shock, one from pulmonary infection, with two of a cause unrelated to stenting (upper gastrointestinal hemorrhage). No other severe EB-RFA-related complications, such as pseudoaneurysm, bile duct perforation, bile leak or gallbladder abscess, were observed. The total incidence of cholangitis was 15.3% (25/163 patients), with being comparable in the two groups. Patients who developed symptoms of bacterial cholangitis were treated successfully with antibiotics.

Table 3 Complications and causes of recurrent biliary obstruction in RFA-stent and stent groups

The total median hospital stay was 13 days (IQR 9, 19), and tended to be slightly longer in the RFA-Stent group (16 days [IQR 9, 20]) than in the Stent group (14 days [IQR 8, 18]) but the difference was comparable (P = 0.075).

The causes of recurrent biliary obstruction in the two groups are shown in Table 3. The main reasons for stent occlusion were tumor ingrowth and overgrowth in both groups. The causes of RBO were comparable between the two groups.

Comparison of therapeutic outcomes before and after PSM

The median follow-up duration was 23.0 months (interquartile range12–29 months) for all patients. Eleven (12.2%) of the 90 patients in RFA-Stent group and 10 (13.7%) of the 73 in Stent group remained alive during the observation period. The median PSP and SSP were 8.0 months (95% CI: 7.4–8.6) and 9.8 months (95% CI: 6.8–12.8) in RFA-Stent group and 5.1 months (95% CI: 4.2–6.0) and 5.1 months (95% CI: 4.1–6.1) in Stent group (P = 0.003, P < 0.001, respectively) (Fig. 3a, b). The mean OS was 8.0 ± 0.6 months and the median OS was 5.7 months (95% CI: 5.1–6.3). Median OS was better in RFA-Stent group than in Stent group (7.0 months (95% CI: 5.5–8.5) vs. 4.5 months (95% CI:3.6–5.4); P = 0.034) (Fig. 4a).

Fig. 3

Primary and secondary stent patency curves of patients who underwent RFA-Stent or Stent for malignant biliary obstruction before and after matching. Primary stent patency curves before (a) and after (c) propensity score matching. PSP There were significant differences in primary stent patency outcomes both before and after propensity score matching (before matching: 8.0 vs. 5.1 months, respectively, P = 0.003; after matching: 8.5 vs. 5.1 months, P < 0.001). Secondary stent patency curves before (b) and after (d) propensity score matching. There were significant differences in primary stent patency outcomes both before and after propensity score matching (before matching: 9.8 vs. 5.1 months, respectively, P < 0.001; after matching: 11.0 vs. 6.0 months, P < 0.001)

Fig. 4

Overall survival curves of patients who underwent RFA-Stent or Stent for malignant biliary obstruction before and after matching. Overall survival (OS) curves before (a) and after (b) propensity score matching. There were significant differences in OS outcomes before and after propensity score matching (before matching: 7.0 vs. 4.5 months, respectively, P = 0.034; after matching: 8.0 vs 4.0 months, P = 0.007)

Multivariable stratified analysis of PSP was performed in all subgroups of patients (Fig. 5). The RFA-Stent provided significant PSP benefits in the subgroups of patients who were male (P = 0.005), aged younger than 65 years (P = 0.009), or had a PS score of 0 (P = 0.037), pancreatic carcinoma (P = 0.031), cholangiocarcinoma (P = 0.035) low-level biliary obstruction (P = 0.01), an initial stricture length > 4 cm (P = 0.008), previous cholangitis (P = 0.004), serum total bilirubin > 165.4 µmol/L (P = 0.001), or serum GGT > 307 U/L (P = 0.002). These results further confirmed the association between RFA-Stent and a decreased stent obstruction risk in patients with MBO.

Fig. 5

Forest plot of primary stent patency in subgroups of patients treated with RFA-Stent and Stent

In the matched cohort, the mean OS was 7.9 ± 0.7 months and the median OS was 5.7 months (95% CI: 4.5 – 6.9) for all 54 pairs of patients. The RFA-Stent group had longer median PSP (8.5 months (95% CI: 7.9–9.1) vs. 5.1 months (95% CI: 4.1–6.1); P < 0.001), SSP (11.0 months (95% CI: 6.8–12.8) vs. 6.0 months (95% CI:5.2–6.8); P < 0.001) and OS (8.0 months (95% CI 6.9–9.1) vs. 4.0 months (95% CI: 2.8–5.2); P = 0.007) than the Stent group (Figs. 3c, d, 4b).

On univariate and multivariate analysis of all patients, only RFA-Stent (P = 0.005) was significantly associated with PSP, and RFA-Stent (P < 0.001) and Serum TB > 165.4 µmol/L (P = 0.015) were independent prognostic factors for PSP (Table 4). RFA-Stent (P < 0.001) and PS status (P = 0.045) were significantly associated with OS, and RFA-Stent (P = 0.008), PS status (P = 0.018) and combination therapy after stent (P = 0.023) were significant independent prognostic factors for OS (Table 5). RFA-Stent (P < 0.001) and female (P = 0.023) were significantly associated with SSP, while RFA-Stent (P < 0.001) and Serum TB > 165.4 µmol/L (P = 0.003) were significant independent prognostic factors for SSP (Supplemental Table S1).

Table 4 Univariate and multivariate analyses of prognostic factors for primary stent patency

Table 5 Univariate and multivariable cox proportional hazards model analysis for risk of death

Discussion

In this study, percutaneous RFA-Stent provided stent patency and OS benefits in the whole population before and after PSM. Also, the results showed percutaneous RFA-Stent had a longer PSP than Stent, especially in patients with unresectable MBO involving low-level biliary obstruction, cholangiocarcinoma, pancreatic carcinoma by subset analysis using a Cox regression model, which can estimate the differential effect of treatment on the tumor characteristics of a study population. Moreover, the complications caused by addition of EB-RFA were tolerable. To the best of our knowledge, our study is the first to evaluate the outcomes of RFA-Stent versus Stent in unresectable MBO through a percutaneous approach using a PSM analysis. In our study, we obtained relatively reliable evidence as to the superiority of percutaneous RFA-Stent over Stent in terms of stent patency and OS in treating MBO. This finding was substantiated by subgroup stratification and propensity score analyses on the basis of statistical methods for controlling of confounders as well as “traditional” multivariable Cox regression analysis.

In our study, percutaneous EB-RFA improved the stent patency in unresectable MBO, especially in some subgroup selected population. Previous studies comparing RFA plus stent insertion with stent implantation alone by endoscopic approach showed inconsistent results; predominantly, these studies demonstrated that a combination of RFA and stent implantation did not provide significant prolongation of stent patency [4, 17, 18], whereas two recent trials showed significant longer stent patency [19, 20]. These differences from our results are most likely attributable to heterogeneity of patients involving in the study, as well as using different material stents and EB-RFA instruments. In comparison, since M. Mizandari et al. first reported the safety and feasibility of percutaneous biliary ablation for MBO [21], previous six retrospective comparative studies in a review article and one randomized controlled trial comparing percutaneous RFA-Stent with Stent showed consistent results [15, 22]. Moreover, our results were confirmed by five meta-analyses involving both endoscopic and percutaneous approach [1, 5, 23,24,25]. The local thermal effect of EB-RFA can significantly increase stricture diameter, which is considered a measure of the effectiveness of EB-RFA in prolonging stent patency [22]. EB-RFA can also be used to clear an occluded metal stent and keep stent function without placement of a new stent [26]. Repeated EB-RFA treatments and multiple sessions were mainly associated with longer stent patency [14], which was similar to a prolongation of SSP in our study. Long stent patency can provide an opportunity for subsequent antitumor therapy in patients with MBO, which can be explained by the result that combination therapy after stent insertion was an independent predictor of survival in multivariable analysis in our study.

Furthermore, Kallis et al. [27] found endoscopic EB-RFA to be a safe and efficacious adjunctive treatment in patients with MBO caused by advanced pancreatic malignancy and suggested that it may provide an early survival benefit. The results were consistent with ours. In our previous study [14], we found that RFA-Stent conferred stent patency benefits in patients with MBO when compared with stenting alone and a trend (albeit not statistically significant) toward a survival benefit with combined treatment. Notably, there were higher proportions of cholangiocarcinoma and younger patients in RFA-Stent group and higher proportions of metastatic diseases in Stent group, as well as a short follow-up period. In the present study, although more patients were included before matching than in the previous study, we found that RFA-Stent had longer PSP, OS and SSP compared with Stent in patients with unresectable MBO and a relative long follow-up period; however, potential confounding factors remained. Therefore, in this retrospective analysis, we conducted a PSM analysis to minimize the potential confounding bias inherent. In the propensity model, where the RFA-Stent and Stent groups had comparable baseline features, RFA-Stent achieved significant improvements in OS and stent patency. The demographics and tumor characteristics of the patients in these studies who underwent RFA-Stent are intrinsically different from those who underwent Stent. Local thermal effect of EB-RFA was mainly associated with the difference of therapy benefits due to different etiology and biological behavior of the tumor [28]. Tumors originating from the bile duct epithelium may have a different response profile from those involving or compressing the bile duct. Well-designed randomized controlled trials with adequate statistical power may be helpful to explore the additional benefits of EB-RFA. So far, only one prospective randomized trial involving 76 patients with MBO has compared RFA-Stent using a percutaneous approach with Stent [15]. The results indicated that percutaneous EB-RFA combined with metal stent insertion can significantly prolong the stent patency without increasing the incidence of complications only in patients with cholangiocarcinoma. Interestingly, the finding in that trial showed no improvement in survival. Notably, all patients in the above prospective trial received brachytherapy procedure, which may obscure the benefits of additional ablation. Moreover, more patients with literally better prognosis (low level obstruction compared with high level obstruction) were involved in Stent group, which also counteracted the benefits of EB-RFA in the combination group.

As previously reported [14, 27, 29], we found complication rates between RFA-Stent and Stent group were comparble. However, severe complications of EB-RFA, such as arrhythmia, bile perforation, and partial liver infarction, can be avoided by adequate pre-procedure evaluation, ablation without overlapping, and careful use of the balloon assistant [14, 30, 31].

There are some limitations in our study. The first limitation is its retrospective design, which unavoidably introduced a degree of selection bias; however, a propensity analysis was performed to minimize the effect of confounding factors. Besides, our study involved different types of etiologies patients with a wide range of naturally different prognoses; So the results should be interpreted with caution, particularly the OS benefits of additional ablation. The second limitation is that potential unmeasured confounders might influence the results due to our study’s retrospective nature with 5 years. For example, users of this technology go through learning curves, which maybe influence the outcomes even with the same user and the same equipment. Thirdly, not all diagnoses of patients were pathologically confirmed in this study. Finally, exploring the molecular mechanism of RFA to bring benefits can increase the credibility of the research results, while we did not examine tumor cell biology or immune regulation change in our study.

Conclusions

In conclusion, our results showed that addition of percutaneous EB-RFA to established stenting therapy has a demonstrable effect in patients with MBO in both before and after PSM analysis. As an initial treatment of MBO, percutaneous RFA-Stent may offer stent patency and survival advantages over Stent, especially in subgroups of patients with favorable characteristics. Large multicenter prospective randomized controlled trials are now needed to confirm the additional benefits of percutaneous RFA-Stent in MBO.