Endoscopic retrograde cholangiopancreatography (ERCP) has been a standard treatment modality for biliary drainage (BD) in cases of malignant hilar biliary obstruction (MHBO). However, ERCP is not always successful due to surgically altered anatomy and duodenal strictures or obstructions.^1,2 For these patients, non-surgical drainage can be an alternative to ERCP. Two drainage methods have been established: percutaneous transhepatic biliary drainage (PTBD) and endoscopic ultrasound (EUS)-guided hepaticogastrostomy (HGS).³ Since EUS-HGS was first presented in 2003 by Burmester et al,⁴ it has been increasingly applied to cases of failed ERCP. Several recent studies^5–8 have demonstrated that EUS-HGS offers excellent technical success (~90%) and reasonable clinical success with an acceptable risk of adverse events such as pancreatitis, tumor ingrowth, or stent malfunction for MHBO. However, the results of EUS-HGS in patients with MHBO are limited, and it is only recommended by experts in EUS interventions, even though EUS-HGS is rapidly expanding and many technological innovations continue to be introduced. This review article focuses on the current status, advantages, and disadvantages of EUS-HGS and discusses future prospects for MHBO based on the currently available literature.

In MHBO, achieving adequate drainage can be frustrating because intrahepatic biliary obstruction can be complex and multiple, and inadequate drainage may result in life-threatening secondary cholangitis. The Bismuth–Corlette classification is generally used, in which Bismuth types III-IV may involve intrahepatic obstruction in more than one segment.⁹ In those cases, achieving adequate BD is not straightforward, since it often requires drainage of more than one segment of the liver; thus, more than one BD procedure may be needed. Moreover, gauging the degree of biliary obstruction is often difficult in real-life clinical practice, and it is especially challenging to predict the adequacy of drainage. Generally, drainage of more than 33% of the total liver volume may be sufficient in patients with preserved liver function, while drainage of more than 50% is generally needed in patients with impaired liver function.¹⁰ Based on this concept, a recent study reported that drainage of more than 50% of the liver measured by a pre-ERCP assessment of the hepatic volume distribution on cross-sectional imaging was a significant factor associated with effective drainage and even longer survival.¹¹ Similarly, Lee et al¹² recently reported that bilateral drainage through the placement of metal stents for inoperable high-grade MHBO was better than unilateral drainage regarding clinical success, median stent patency, and the need for re-intervention. In a comparative study between metal and plastic stents, Xia et al¹³ found that the overall clinical success rate was 98.9%, 83.5%, 71.4%, and 65.4% in the bilateral metal, unilateral metal, bilateral plastic, and unilateral plastic groups, respectively. Therefore, they concluded that bilateral metal stent placement was better than unilateral placement regarding stent patency, the need for reintervention, and even overall survival. In terms of the type of metal stent placement, Naitoh et al¹⁴ demonstrated that side-by-side (SBS) placement was associated with better stent patency than stent-in-stent (SIS) placement, but with an increased risk of adverse events. Contrary to this, Lee et al¹⁵ demonstrated that SIS placement showed similar clinical efficacy and adverse event rates compared to SBS placement, with a marginally significant difference in stent patency at 3 months and no statistically significant difference at 6 months. Since there is still no consensus regarding the best strategy for effective drainage in MHBO and the endoscopic approach in patients with complex anatomy is technically challenging, other alternatives such as PTBD, EUS-guided drainage, or a multimodal approach with the addition of other modalities should be applied.

Since EUS-BD was first reported in 2001,¹⁶ it has become an effective strategy for the management of malignant biliary obstruction (MBO) in patients who have experienced unsuccessful ERCP or have an inaccessible papilla due to surgically altered anatomy and duodenal strictures or obstructions.¹ The EUS-BD procedures are classified as follows: (1) EUS-HGS, (2) EUS-guided choledochoduodenostomy (EUS-CDS), (3) EUS-guided hepaticoduodenostomy (EUS-HDS), (4) EUS-guided rendezvous procedure (EUS-RV), and (5) EUS-guided anterograde stenting (EUS-AS).

However, the pathophysiological mechanism of advanced MHBO, including Bismuth type III or IV, is different from that of distal MBO.¹⁷ In cases of distal MBO including Bismuth type I or II, drainage with a single stent can generally be sufficient because it can be responsive to the majority of liver parenchyma while multiple stent placement should theoretically be required to achieve adequate drainage or prevent focal cholangitis in advanced MHBO. In a study regarding EUS-BD in patients with MHBO, Moryoussef et al⁸ demonstrated that clinical success on days 7 and 30 after EUS-HGS was 72.2% and 68.8%, respectively, although the technical success rate was 94%. We focused on EUS-HGS with transmural stenting because the passage of a transhepatically placed guidewire into the duodenum across high-grade MHBO is technically challenging.¹⁸

Basic technique of EUS-guided biliary drainage in malignant hilar biliary obstruction

The selection of an adequate target intrahepatic duct (IHD) for EUS-HGS is the most important technical step. Generally, B2 can be accessed from the gastroesophageal junction or cardia while B3 can be accessed from the lesser curve of the stomach body.¹⁹ Considering the direction of the IHD, the approach to B2 may be more ideal for EUS-RV or EUS-AS because B2 may be connected straightly with the hilum or even the common bile duct. In contrary to this, the approach to B3 may be more suitable for EUS-HGS with transmural stenting.²⁰ In rare cases with surgically altered anatomy, such as total gastrectomy or hepaticoenterostomy (e.g., hepaticoesophagostomy, hepaticojejunostomy, or hepaticoduodenostomy) can be applied.

To ensure the appropriate approach to the dilated IHD, a 19-gauge standard aspiration needle, which is unmounted with a stylet, can be used to create a puncture at either the lesser curvature or the cardia of the stomach. A contrast medium is then injected to visualize the bile duct, including the obstructive IHD of MHBO under fluoroscopy. A 0.025-inch angle tip guidewire is passed through the needle into the targeted IHD. Every attempt should be made to pass the guidewire into the small bowel across the biliary stricture for stabilization of the guidewire. Where this is not possible, the guidewire should be coiled at least in the liver hilum or the opposite IHD for successful transmural stent placement. After withdrawal of the needle, an ultra-tapered ERCP catheter can be inserted over the guidewire for dilation of the transmural tract. Then, a balloon dilation catheter can be inserted over the guidewire and dilation should be performed for each segment, including the IHD, the hepatic parenchyma, and the stomach wall. At the discretion of the endoscopist, a balloon catheter can be used for initial dilation without preceding dilatation using the ERCP catheter. If resistance to advancement of the ERCP catheter or balloon catheter is noted, a needle-type knife or cystotome can be used with electric current. After sufficient dilation of the tract, a partially or fully covered expandable metal stent especially designed for EUS-BD, or a plastic stent, is then transgastrically inserted over the guidewire into the left IHD.²¹ In a metal stent, the length is determined by approximating the distance between the IHD and the stomach, with the addition of extra length (approximately 20 mm) (Fig. 1).²² After positioning the delivery system within the IHD through a guidewire, the stent deployment hub is released to deploy the distal portion of the stent. The echoendoscope is then retracted to visualize the catheter shaft in the stomach, and the catheter deployment hub is then completely released to deploy the proximal flange within the gastric lumen. During the placement of the plastic stent, balloon dilation should be applied to sufficiently dilate the tract before the plastic stent is deployed into the target IHD (Fig. 2). The length of the stent is also determined by approximating the distance between the IHD and the stomach with the addition of extra length (approximately 20 mm) at both sides. To prevent inner migration of the plastic stent, the stent should have a pigtail configuration at the proximal end. Table 1 summarizes the possible stent types, as well as the advantages and disadvantages of each stent for EUS-guided peripancreatic fluid collection drainage.

Table 1 . Comparison of the Characteristics of Each Stent for Endoscopic Ultrasound (EUS)-Guided Bile Duct Drainage.

Type of stent	Possible type	Advantage	Disadvantage
Plastic stent	- Single-pigtail stent - Double-pigtail stent	- Less expensive - Convenience for revision - Less shortening of stent - Less tumor tissue hyperplasia or ingrowth/overgrowth	- Frequent stent dysfunction including occlusion due to the small diameter of the stent - Possible bile leakage along the stent
Fully or partially covered self-expandable metal stent (FCSEMS or PCSEMS)	- PCSEMS or FCSEMS without anti-migrating flanges - PCSEMS or FCSEMS with anti-migrating flanges (unidirectional or bidirectional)	- Longer stent patency than plastic stent due to larger diameter of the stent - Theoretically less bile leakage	- Relatively expensive - Side-branch obstruction leading to biloma or hepatic abscess - Stent malposition - Higher rate of stent shortening leading to stent migration
Lumen-apposing metal stent (LAMS)	- Only for EUS-guided choledochoduodenostomy - LAMS with bidirectional anchoring flanges but no electrocautery-enhanced tip - LAMS with electrocautery-enhanced tip	- Largest diameter - Longer stent patency - Shorter procedure time due to one-step creation of the fistula - Omitting unnecessary steps, including balloon dilation - Migration is rare	- Very expensive - Rare but possible adverse events of significant bleeding or buried LAMS syndrome

Figure 1. Endoscopic ultrasound (EUS)-guided hepaticogastrostomy (HGS) using a dedicated partially covered self-expandable metal stent (PCSEMS) (hybrid metal stents; Standard Sci Tech Inc., Seoul, Korea) for obstruction of the left intrahepatic duct (IHD) in malignant hilar biliary obstruction. (A) After failed endoscopic retrograde cholangiopancreatography, EUS-HGS was performed in patients with left intrahepatic obstruction. On EUS, a dilated left IHD was observed on the expected needle track. (B, C) The left IHD (B3) was punctured by a 19-gauge standard needle. (D) The guidewire was delivered up to the main left IHD. (E) The fistula tract was dilated using a 4-mm balloon catheter. (F) A PCSEMS with anchoring flaps was placed through hepaticogastrostomy. Fluoroscopic image showing the stent placed between the stomach and the left IHD.

Figure 2. Endoscopic ultrasound (EUS)-guided hepaticogastrostomy (HGS) using a 7-Fr double-pigtail plastic stent (DPPS) for obstruction of the left intrahepatic duct (IHD). (A) After failed cannulation to the left IHD, a dilated left IHD was observed on the expected needle track for EUS-HGS. (B, C) The left IHD (B3) was punctured by a 19-gauge standard needle. (D) The fistula tract was dilated using a 4-mm balloon catheter. (E) Fluoroscopic image showing the 7-Fr DPPS placed between the stomach and the left IHD. (F) Endoscopic image showing the drainage of pus-like material through the 7-Fr DPPS.

Technical challenges of EUS-guided biliary drainage in malignant hilar biliary obstruction

In transmural stenting in B3, the endoscopist should withdraw the echoendoscope for visual identification of the distal end of the stent. However, at this moment, the stent may eventually migrate to a more inner side of the IHD during stent deployment, especially in patients with a J-shaped stomach.^19,23 To prevent inner migration after stent deployment, a stent longer than 8 cm is recommended.¹⁹ Another important tip for the technique of appropriate stent deployment is that the echoendoscope should be constantly placed at the initial puncture site and attached to the gastric wall.¹⁹ Holding on the scope position, all procedures including tract dilation and even stent deployment should be performed under EUS and fluoroscopic, not endoscopic guidance, to prevent accidental falling out of guidewire or stent migration. If the deployed metal stent appears to be relatively short for the transmural tract, an additional metal stent or plastic stent with a double-pigtail configuration can be considered in a stent-in-stent manner with adequate length as a bridge between the liver parenchyma and stomach wall.²⁴

Clinical outcomes of EUS-guided biliary drainage in malignant hilar biliary obstruction

Although several case reports or observational studies for EUS-BD in MHBO are available, large-scale randomized controlled trials (RCTs) or comprehensive meta-analyses are not reported. A recent retrospective study by Winkler et al²⁵ demonstrated that the technical success rate for EUS-HGS with transmural drainage of MHBO was 100%, and the clinical success rate was 95% among 16 patients. Another study by Bories et al⁵ that analyzed 11 patients with failed ERCP for MHBO demonstrated that 10 patients successfully underwent EUS-HGS, using a plastic stent in seven patients and a metal stent in three patients. In addition, a retrospective analysis²⁶ of 30 patients with MHBO of Bismuth II or higher grade who underwent EUS-HGS after failed ERCP reported that technical success with EUS-HGS was achieved in 29 of 30 patients (96.7%) and clinical success was attained in 22 of these 29 (75.9%). Regarding adverse events, bile peritonitis occurred in three patients (10.0%) and stent dysfunction occurred in seven patients (23.3%) although there was no procedure-related mortality. Furthermore, Bismuth IV stricture was found to be a strong predictor of clinical failure (odds ratio [OR] = 12.7; 95% confidence interval [CI], 1.18–135.4; P = 0.035) in the multivariate analysis. In a recent comparative study²⁷ between endoscopic and percutaneous approaches for patients with Bismuth–Corlette class III, IV, or more advanced MHBO, the technical and clinical success rate was 84.2 % (16/19) and 78.9 % (15/19) in the endoscopic drainage group (ERCP + EUS-HGS) and 100 % (17/17; P = 0.23) and 76.5 % (13/17; P > 0.99) in the bilateral PTBD group, respectively. Furthermore, the overall adverse event rate was 26.3 % (5/19) in the endoscopic drainage group and 35.3 % (6/17; P = 0.56) in the bilateral PTBD group. The recurrent biliary obstruction (RBO) rate within 3 months was 26.7 % (4/15) and 88.2 % (15/17; P = 0.001) in the endoscopic drainage group and the bilateral PTBD group, respectively, while the RBO rate within 6 months was 22.2 % (2/9) vs. 100 % (9/9; P = 0.002), respectively. The median number of revisions for RBO at 3 months (0 [interquartile range, 0–1]) was significantly lower in the endoscopic drainage group than in the bilateral PTBD group (1 [interquartile range, 1–2.5], P < 0.001). In addition, the median time to RBO was significantly longer in the endoscopic drainage group than in the bilateral PTBD group (92 [56–217] vs. 40 [13.5–57.8] days, respectively; P = 0.06).

New techniques and concepts for EUS-guided biliary drainage in right intrahepatic duct obstruction

In right-sided intrahepatic biliary obstruction or Bismuth–Corlette class III, IV or more advanced MHBO, EUS-HDS may be suitable instead of EUS-BD from the stomach (EUS-HGS).²⁸ In the first trial reporting EUS-HDS for isolated right IHD obstruction by Park et al,²⁹ technical success was achieved in five of six patients. After EUS-guided cholangiography of the right IHD, they conducted three kinds of sequential procedures: (1) using a cholangiography by EUS-guided transduodenal puncture as a “roadmap” to facilitate super-selection for the right IHD by ERCP, (2) antegrade stenting or balloon dilation after EUS-guided transduodenal puncture in the right main IHD, and (3) transmural stenting between right IHD and duodenum in cases of failure of guidewire delivery to the hilum (Fig. 3). However, EUS-HDS should be conducted in limited cases of highly selected patients by experienced interventional echoendoscopists because the right IHD may occasionally not be adjacent to the duodenum or the portal vein could be close to the tract of approach. Furthermore, optimal visualization, successful puncture, or delivery of the guidewire for the right IHD is often difficult even in those with marked right IHD dilatation.¹⁸ To overcome these limitations, echoendoscopists can make a position of echoendoscope in a “U” shape under fluoroscopic guidance because the optimal site and nearest distance between the echoendoscope and right IHD can be achieved in this position.

Figure 3. Endoscopic ultrasound (EUS)-guided hepaticoduodenostomy (HDS) using a 7-Fr double-pigtail plastic stent (DPPS) for obstruction of the right intrahepatic duct (IHD). (A) After failed cannulation to the right IHD, EUS-HDS was performed in patients with right intrahepatic obstruction. On EUS, a dilated right IHD was observed on the expected needle track. (B, C) The right posterior IHD was punctured by a 19-gauge standard needle. (D) The guidewire was delivered up to only right posterior IHD. (E) The fistula tract was dilated using a 4-mm balloon catheter. (F) A 7-Fr DPPS was placed through hepaticoduodenostomy. A fluoroscopic image showing the stent placed between the duodenum and the right posterior IHD.

Safety of EUS-guided biliary drainage in malignant hilar biliary obstruction

Although previous trials suggested that EUS-BD is feasible and effective, major concerns have been raised regarding the potential risks of serious adverse events, particularly stent-related adverse events. The overall rate of post-procedure adverse events, including stent migration, bile leakage, pneumoperitoneum, and cholangitis, was found to be 8% in patients who underwent EUS-BD in MHBO,³⁰ and the rate in patients with proximal biliary obstruction, including MHBO, was found to be higher than distal biliary obstruction, although it did not reach statistical significance.³¹ Among these, stent migration is a major adverse event that can lead to life-threatening consequences, such as bile leakage with or without bile peritonitis and bile duct obstruction with or without biloma. Several studies^18,32,33 reported a high probability of outward stent migration that could be derived from the difficulty of identifying the distal end of the stent under fluoroscopic guidance and shortening of the stent itself. Furthermore, inward migration of the stent can also lead to bile leakage, obstruction of IHD or biloma, or even cholangitis. Theoretically, plastic stents have a relatively small lumen that leads to frequent stent occlusion and revision, while a self-expanding metal stent has a relatively large diameter, which might enable longer patency.³⁴ Therefore, self-expanding metal stents may be advantageous because they have (1) better drainage efficacy due to a larger diameter than plastic stents, (2) tamponade effects for the prevention of bile leakage or bile peritonitis; and (3) a tamponade effect for preventing tract bleeding due to compression derived from the radial force of the stent.^5,19 However, self-expanding metal stents may have limitations, such as high costs, a shortening rate of about 40% in all braided-type stents, a risk of fatal adverse events (e.g., unexpected migration), and direct obstruction of adjacent IHD.^35,36 Regarding post-procedural bleeding, Park et al. reported that two cases of mild bleeding occurred among 55 patients with EUS-BD, although no results are available for MHBO.²⁴

Comparison of EUS-guided biliary drainage and other modalities in malignant hilar biliary obstruction

Comparing with ERCP for drainage of MHBO, EUS-BD showed similar efficacy with a superior safety profile, and in particular a lower risk of post-procedural pancreatitis. The reported rates of technical success, clinical success, and adverse event of EUS-BD were 93.4%–100%, 90.2%–100%, and 6.3%–15%, while those of ERCP were 90.2%–94.2%, 91.3%–94.5%, and 8.7%–24%, respectively.^37–40 Furthermore, a recent meta-analysis⁴¹ demonstrated comparable clinical success and adverse event rates between EUS-BD and ERCP, with a significantly lower risk of pancreatitis after EUS-BD than after ERCP (0.3% vs. 7.3%), although there were no comparative studies between the two modalities in MHBO. The combination of EUS-BD and ERCP was a novel strategy that appeared to be a feasible alternative to PTBD in the treatment of MHBO with advanced Bismuth type.^17,35 This combination strategy is based on the principle that internal drainage is more physiologic and comfortable than PTBD. A recent multicenter observational study²⁷ for MHBO reported that the rates of technical success, clinical success, and adverse events for the combination strategy and PTBD were 84.2% (16/19) vs. 100% (17/17, P = 0.23), 78.9% (15/19) vs. 76.5% (13/17, P > 0.99), and 26.3% (5/19) vs. 35.3% (6/17, P = 0.56), respectively. The RBO rates within 3 and 6 months of the combination strategy and PTBD were 26.7% (4/15) vs. 88.2% (15/17; P = 0.001) and 22.2% (2/9) vs. 100% (9/9; P = 0.002), respectively.

In comparing EUS-BD and PTBD in the treatment of MHBO, several studies^42–46 reported that EUS-BD showed comparable technical (86.4%–100%) and clinical success (62.2%–100%) rates, with a lower rate of adverse events (6.6%–15.3%). In addition, a recent RCT demonstrated that EUS-BD and PTBD had similar efficacy in patients with unresectable MHBO and inaccessible papilla based on rates of technical and functional success and quality of life, although fewer adverse events and unscheduled re-interventions were reported in EUS-BD.⁴⁶ In a recent meta-analysis⁴⁷ with nine studies for a total of 483 patients, no difference was found in technical success between EUS-BD and PTBD (OR = 1.78; 95% CI, 0.69–4.59) although EUS-BD showed better clinical success (OR = 0.45; 95% CI, 0.23–0.89), fewer post-procedure adverse events (OR = 0.23; 95% CI, 0.12–0.47), and a lower rate of reintervention (OR = 0.13; 95% CI, 0.07–0.24).

Given the various treatment options for MHBO, EUS-BD can be a promising modality and reasonable alternative to PTBD after failed ERCP because it is reliable, minimally invasive, and safe. However, these endoscopic techniques have not yet been established as standard procedures and their roles are limited to treating patients in whom ERCP has failed. Furthermore, although no conclusion has been reached regarding which approach is preferred, these procedures should be commonly considered by endosonographers who are skilled in ERCP and interventional EUS at high-volume institutions. Further developmental innovations and technical refinements for EUS-BD may be warranted for the generalization of this procedure.

None.

No potential conflict of interest relevant to this article was reported.