Therapeutic endoscopic ultrasound

INTRODUCTION — Endoscopic ultrasound (EUS) was developed as a diagnostic modality but rapidly gained a role for a variety of therapeutic applications. EUS has been used increasingly for drainage of pancreatic fluid collections, treatment of cystic lesions of the pancreas, EUS-guided cholangiopancreatography, localized therapy for pancreatic tumors, and treatment of gastric varices.

This topic will provide an overview of clinical applications of therapeutic EUS. The indications, patient preparation, technical aspects, and adverse events associated with EUS examination of the upper gastrointestinal tract are discussed separately. (See "Endoscopic ultrasound: Examination of the upper gastrointestinal tract".)

DRAINAGE PROCEDURES — EUS is well-suited to safely drain fluid collections of various types in areas accessible from the stomach, duodenum, or rectum. The most abundant experience has been with drainage of pancreatic fluid collections, but case series have described other drainage procedures (eg, abscesses, bilomas) [1-11].

Pancreatic fluid collections — During EUS-guided drainage of a pancreatic fluid collection, the fluid cavity is accessed via the creation of a tract through the gastric or duodenal wall with subsequent stent placement.

The EUS-guided drainage of walled-off pancreatic fluid collections is discussed in detail separately:

●(See "Approach to walled-off pancreatic fluid collections in adults".)

●(See "Endoscopic interventions for walled-off pancreatic fluid collections".)

Postoperative fluid collections — Abdominal postoperative fluid collections (POFC) are often drained percutaneously, while data suggest that EUS-guided drainage and stent placement is a reasonable alternative [12-15]. However, the optimal timing of EUS-guided intervention is uncertain. In a retrospective study of 75 patients with symptomatic postoperative fluid collections (due to distal pancreatectomy and splenectomy in 63 patients [84 percent]), clinical success rates (ie, resolution of symptoms and fluid collection) were not significantly different for patients who had early drainage (≤30 days after surgery) compared with late drainage (>30 days after surgery) (93 versus 94 percent) [16]. In addition, rates of adverse events were not significantly different for patients who underwent early EUS-guided drainage compared with those undergoing late drainage (21 versus 30 percent). Whereas endoscopic drainage is often delayed (ie, ≥4 weeks) until a well-defined, thick wall has formed around POFC to optimize safety, an earlier EUS-guided intervention may be feasible if these results are validated by other studies. (See "Surgical resection of lesions of the body and tail of the pancreas", section on 'Perioperative morbidity and mortality'.)

Liver abscess — EUS-guided drainage and stent placement has been described for treating liver abscesses that were difficult to access via percutaneous technique [17-20]. Early studies involved using nasocystic drainage tubes and double-pigtail stents, with subsequent reports introducing the use of fully covered metal stents or a lumen-apposing metal stent (LAMS) [17-19,21]. In a review of 15 case reports and case series including 39 patients with liver abscess, EUS-guided drainage was associated with technical success in 98 percent of patients [20]. While an EUS-guided approach is less invasive than surgical drainage, EUS relies on access from the upper gastrointestinal tract, and thus, EUS drainage is typically limited to treating collections in the left lateral and caudate lobes. Additionally, this is a technically complex procedure requiring advanced endoscopic expertise, which may not be widely available.

Bilomas — Case reports have described successful EUS-guided drainage of bilomas [6-8]. The largest report described four patients with bile leaks and gallbladder fossa bilomas who were drained via EUS-guided double pigtail stent placement [6]. ERCP was performed initially with placement of a biliary endoprosthesis. Subsequently, a 19-gauge needle was used to access the biloma from the stomach or proximal duodenum, a 0.035 inch guidewire was used to secure access, and 6 to 8 mm balloon dilators were used to dilate the fistula. In one case, a single pigtail stent was placed, and in the other three, one to two double pigtail stents were left in place to drain the bilomas. The size of the fluid collections ranged from 4 x 2 cm to 8 x 6 cm. Antibiotic coverage was given the day of the procedure and for 10 to 14 days afterward. The duration of drainage was 11 days in one case, and one to three months in the others. No complications were reported.

Gallbladder — Cholecystectomy is generally preferred for management of acute cholecystitis but is not always an option due to patient comorbidities. Case series have demonstrated the feasibility of EUS-guided gallbladder drainage in such patients. This is discussed in detail separately. (See "Treatment of acute calculous cholecystitis", section on 'Endoscopic'.)

Pelvic abscess — Data from case series suggest that EUS-guided drainage for treating pelvic abscess was effective and safe [3-5,22]. Stents (with or without drainage catheters) were used in the initial EUS studies, but data on single-step procedures using a lavage and antibiotic instillation technique are emerging.

●Stent placement – One series involved 25 patients with pelvic abscesses that were predominantly postsurgical or posttraumatic [5]. Patients with abscesses less than 8 cm in size were treated with two 7 Fr transrectal stents. If the abscess was 8 cm or more in size, a 10 Fr drainage catheter was also placed. The drainage catheter was removed once the abscess had decreased in size by at least 50 percent, and the stents were removed after two weeks. The abscesses were successfully drained in all patients at the time of EUS, and there were no procedure-related complications. For 24 patients, there was complete resolution of the abscesses on follow-up imaging at two weeks. Another case series included 12 patients with abscesses mostly from pelvic surgery [4]. Stents were inserted successfully in nine patients, resulting in complete resolution of the abscesses in eight patients, and incomplete resolution in one patient with a large cyst (>8 cm) that required subsequent surgical management. Stents were removed after three to six months. No major complications were reported.

●EUS-guided lavage with antibiotic instillation – Limited data suggest that single-procedure abscess drainage by replacing the purulent collection with dilute gentamicin is technically feasible and effective. In a case series including six patients with pelvic abscess, transrectal EUS-guided abscess drainage using a 19-gauge needle was initially performed and followed by serial lavage with sterile saline until the cavity was cleared of pus. These maneuvers were followed by instillation of gentamicin 40 mg/mL [22]. Complete abscess resolution was achieved in four patients, while the remaining two patients had improvement but without complete resolution.

FINE-NEEDLE INJECTION — EUS permits precise targeting for the delivery of various substances directly into pancreas, liver, or subepithelial lesions. EUS-guided fine-needle injection (FNI) of various agents has also been reported for treating insulinomas [23], hepatic metastases [24], esophageal cancer [25,26], cystic neoplasms of the pancreas [27,28], and pancreatic adenocarcinoma [29,30].

Pancreatic cystic neoplasms — The role of ablative therapies administered by EUS for the treatment of pancreatic cystic neoplasms is discussed elsewhere. (See "Pancreatic cystic neoplasms: Clinical manifestations, diagnosis, and management", section on 'Management'.)

Insulinoma — A case report described EUS-FNI of a symptomatic 13 mm insulinoma [23]. The tumor was ablated after injecting 8 cc of 95 percent ethanol in four 2 mL aliquots. No tumor recurrence was observed based upon surveillance EUS or development of recurrent symptoms after 34 months of follow-up.

Hepatic metastases — Several studies have described percutaneous ablation of hepatic metastases with ethanol injection. Ablation with EUS-FNI has been described only in a case report [24]. The patient had a solitary liver metastasis from rectal cancer, measuring 1.6 cm that was not amenable to percutaneous approach due to vascular interposition. Ethanol, at a concentration of 95 percent and a volume averaging 6 mL, was used during multiple EUS-FNI sessions several weeks apart.

The patient had a good response, with disappearance of the lesion and a decrease in CEA level. He eventually had an additional lesion that was treated similarly with four sessions with a good response. The only complication was a subcapsular hematoma after one procedure. At the time of the report, the patient was still alive and in relatively good health 5.5 years after the initial diagnosis of metastatic cancer.

Pancreatic adenocarcinoma — Various biologic anti-tumor agents have been introduced into pancreatic cancers under EUS-guided FNI for control of locally advanced disease. Although long-term results are not well studied, preliminary results suggest these approaches are generally safe and may prove to be an adjunct or alternative to traditional chemoradiation therapies:

●Cytoimplant – EUS-guided injection for the treatment of advanced pancreatic cancer has been reported [29]. In a phase I study including eight patients with pancreatic cancer, a mixed lymphocyte culture of donor and host mononuclear cells (cytoimplant) was injected into the tumor with EUS-guided FNI. Two patients had partial responses and one had a minor response with a median survival of 13 months. No procedure-related complications occurred.

●TNFerade – TNFerade is a replication-deficient adenoviral vector that contains the human TNF-alpha gene regulated by Egr-1, a chemoradiation inducible promoter. When combined with subsequent chemoradiation, TNF-alpha levels in the tumor increase, potentially leading to tumor suppression [31].

A multicenter study described transfer of TNFerade into locally advanced pancreatic cancer via endoscopic ultrasound or percutaneous guidance [32]. Of the 37 patients treated, stabilization of tumor was observed in 83 percent at one month and 74 percent at three months. Tumor area reduction of greater than 25 percent at one month and 50 percent at three months occurred in 31 and 11 percent of cases, respectively. Survival without overall progression was 63 percent at one month and 47 percent at three months. The therapy was generally well tolerated and considered safe and effective.

A phase I/II study found less promising results [33]. TNFerade was injected into locally advanced pancreatic carcinomas in 50 patients. Only 20 patients (40 percent) were free from local progression at three months.

In 2010, a phase III trial using TNFerade injections into locally advanced pancreatic adenocarcinoma was terminated early when interim analysis showed only an 8 percent lower risk of death with projected inability to reach a statistically significant improvement in outcome.

Brachytherapy — Interstitial brachytherapy is used for treatment of various cancers including head, neck, breast, lung, prostate, and gastrointestinal tract. After radioactive seed placement, the target tissue is exposed to gamma radiation, producing localized tissue injury and tumor ablation. A potential advantage of brachytherapy over traditional external beam radiotherapy is limited radiation toxicity to surrounding normal tissue. Radiation therapy techniques are discussed separately. (See "Radiation therapy techniques in cancer treatment", section on 'Brachytherapy'.)

For brachytherapy, radioactive seeds are usually implanted under CT-guidance or during laparotomy. However, EUS-guided brachytherapy has been reported for head and neck tumors [34], recurrent esophageal cancer with perigastric adenopathy [25], and pancreatic adenocarcinoma [35].

Studies have suggested that EUS-guided brachytherapy is technically feasible and generally well tolerated. However, EUS-guided brachytherapy has failed to gain traction as a standard treatment option. An illustrative series included 15 patients with unresectable pancreatic cancer who underwent EUS-guided implantation of an average of 22 radioactive seeds [30]. During a median follow-up of 11 months, 27 percent of patients demonstrated a partial tumor response, 20 percent showed a minimal response, and 33 percent demonstrated stable disease. Clinical benefit was shown in up to 30 percent of patients, mostly due to a reduction in pain. Pancreatitis occurred in three patients with formation of a pseudocyst in two patients. However, these adverse events were mild and were treated conservatively without additional therapy.

Esophageal cancer — TNFerade is an adenoviral vector that carries the transgene encoding human TNF-alpha, and it has been used for treatment of locally advanced esophageal cancer. A multicenter phase I trial reported experience with 24 patients who had stage II and III esophageal cancer [26]. The vector was injected directly into the tumors, 18 patients with traditional endoscopy and six with EUS-FNI. Results were promising with minimal morbidity, and a 40 percent pathologic complete response in the top three dose cohorts. At the time of the final report, median follow-up was 42.9 months, and three- and five-year disease-free survival rates were 57 and 52 percent, respectively. Further studies have not been published. There was a non-significant trend toward increased efficacy but prolonged procedure times using EUS compared with traditional endoscopy.

A case report described the use of EUS-FNI to place specially adapted brachytherapy beads for nodal spread of esophageal squamous cell carcinoma [25]. A patient with T3 N1 M0 disease initially treated with neoadjuvant chemoradiotherapy followed by surgical esophagectomy had disease recurrence to a celiac lymph node. The node was treated with CT-guided brachytherapy bead placement. Subsequent recurrence manifested by two lymph nodes at the diaphragmatic crus was not amenable to percutaneous placement; as a result, EUS-FNI was used. Two I(125) beads were placed that had been designed to fit within a 19-gauge needle, which was sealed with sterile bone wax to prevent a bead from prematurely dislodging from the needle. No recurrence was detected at eight months follow-up.

Gastrointestinal bleeding — The use of EUS-guided intervention for gastrointestinal bleeding includes:

●Refractory bleeding lesions – EUS guidance was used for treatment by FNI into lesions with at least two failures of hemostasis by conventional endoscopy [36]. The series included one Dieulafoy lesion and one pancreatic pseudoaneurysm successfully treated with ethanol, and one duodenal ulcer and two gastrointestinal stromal tumors treated with cyanoacrylate. No rebleeding was observed.

●Esophageal varices – A controlled trial included 48 patients who were randomly assigned to sclerotherapy of esophageal varices using EUS guidance or using standard endoscopy with ethanolamine as the sclerosant [37]. There was no significant difference in the number of sessions required for obliteration (4.1 versus 4.3) or time to successful obliteration (59 versus 63 days). No difference was observed in variceal recurrence rates.

●Gastric varices – Multiple series have evaluated EUS-guided injection for treatment of gastric varices [38-44]:

•One series of 101 patients used 0.5 mL cyanoacrylate and 0.7 mL lipiodol and compared "on demand" injection after the index procedure (reserved for recurrent bleeding) with scheduled biweekly procedures [39]. Successful obliteration of varices was observed in 80 percent of patients in an average of just over two sessions, spanning approximately five weeks. A median of three doses of cyanoacrylate was required for obliteration. Repeated procedures on a scheduled basis significantly reduced the frequency of late rebleeding.

•A prospective series of five patients with bleeding gastric varices described experience with EUS-guided injection of 1 mL of 1:1 cyanoacrylate-lipiodol or Glubran2-lipiodol, focusing on localization of the perforating veins as the target for injection [38]. Hemostasis was achieved in all five patients. Eradication of the varices was successful in two patients after one session, and in three patients after two sessions (mean 1.6 sessions). There were no cases of recurrent bleeding over a 10-month follow-up period, and no complications were encountered. The authors noted that EUS-FNI may have an advantage over traditional endoscopy due to a lack of need for good endoscopic visualization to target the injection.

•EUS-FNI with cyanoacrylate versus coil application was compared in a study including 30 patients with gastric varices. Gastric varices were obliterated in over 90 percent of patients in both groups. However, cyanoacrylate glue was associated with a higher rate of adverse events such as glue emboli (58 versus 9 percent). The authors concluded that the two modalities were effective but coil injection had a better safety profile [41].

•A series of 152 patients with gastric fundal varices who were treated with combined injection of coils and cyanoacrylate glue reported technical success in 151 patient (99 percent) [42]. The mean number of coils was 1.4 (range 1 to 4 coils) and mean volume of cyanoacrylate was 2 mL (range 0.5 to 6 mL). Among patients who had a follow-up EUS after a mean of 436 days, obliteration of gastric varices was confirmed in 93 percent of patients. Adverse events included pulmonary embolism (one case) and delayed gastrointestinal bleeding from coil or glue extrusion (four cases).

•In a cohort study including 104 patients with gastric fundal varices, EUS-guided FNI with cyanoacrylate glue was compared with direct endoscopic injection [43]. EUS-FNI was associated with a lower rate of rebleeding (9 versus 24 percent), despite using less cyanoacrylate glue (mean volume 2.0 versus 3.3 mL) and injecting a higher number of varices (1.6 versus 1.1 varices) in the EUS group.

Data suggest that combining EUS-guided cyanoacrylate glue plus coil application was more effective than either method alone. In a systematic review of 11 studies including 536 patients, EUS-FNI of cyanoacrylate glue plus coils was associated with higher rates of clinical success compared with either cyanoacrylate glue alone or coils alone (98 versus 96 percent and 96 versus 90 percent, respectively) [44]. In addition, EUS-FNI of cyanoacrylate glue plus coils was associated with lower rates of adverse events compared with cyanoacrylate glue alone (10 versus 21 percent).

Achalasia — The use of EUS-guidance for injection of botulinum toxin into the lower esophageal sphincter has been reported in small case series [45,46]. Botulinum toxin injection for treating achalasia is discussed in more detail separately. (See "Pneumatic dilation and botulinum toxin injection for achalasia", section on 'Botulinum toxin injection'.)

Obesity — EUS-guided botulinum toxin injection into the antrum has been studied for the treatment of obesity in a pilot study, hypothesizing that delayed gastric emptying would result in weight loss. In the open-label study, 10 patients received an EUS-guided injection of 100 or 300 units of botulinum toxin into the muscularis propria of the antrum and were then followed for 16 weeks [47]. The maximum tolerated volume during a nutrient drink test (a measure of satiety) decreased from 1380 mL at baseline to 620 mL two weeks after injection, though the decrease was only statistically significant in those who received 300 units of botulinum toxin. Gastric emptying was not significantly prolonged in patients who received the higher dose. The mean weight loss was 5 kg after 16 weeks.

OTHER CLINICAL APPLICATIONS

EUS-guided tissue ablation — A common clinical dilemma is the management of cystic and neuroendocrine tumors. Given the variability in clinical behavior and morbidity of surgical resection, local therapy would be an attractive option if adequate safety and efficacy data were available. While prospective human trials documenting safety and efficacy of treating pancreatic cystic lesions exist [27,28], solid tissue ablation utilizing EUS-guided ethanol injection or photodynamic therapy catheter insertion in the pancreas has been described only in case series or animal studies [23,48-50]. As a result, use of these techniques is limited to research protocols.

EUS-guided cholangiopancreatography — EUS-guided cholangiopancreatography is a technique for gaining access to the bile ducts and/or the pancreatic duct in patients when endoscopic retrograde cholangiopancreatography (ERCP) has failed or was not possible due to altered surgical anatomy or obstructing duodenal or ampullary tumors [51-55]. Due to their proximity to the stomach and duodenum, the left intrahepatic ducts, common bile duct, and main pancreatic duct are well visualized with EUS and can be accessed with either a transgastric or transduodenal approach.

Accessing the biliary duct — Three EUS-guided approaches have been used to decompress an obstructed biliary system:

●Creation of a duodenobiliary fistula with stent placement.

●Creation of a gastrobiliary fistula with stent placement.

●EUS-guided placement of a guide wire in the common bile duct (by accessing either the left intrahepatic ducts or common bile duct) followed by endoscopic retrograde cholangiography and biliary stent placement (also referred to as a rendezvous procedure) (image 1).

EUS-guided cholangiography with biliary drainage and stent placement has been performed to relieve biliary obstruction [55-59], and this technique offers an alternative to surgical or percutaneous drainage for patients in whom conventional ERCP was unsuccessful or not possible, or for patients who do not want an external drain. However, the risk of complications including bleeding, pancreatitis, perforation, and stent migration remains a concern. EUS-guided biliary drainage should be performed at centers with advanced endoscopists and a multidisciplinary team (ie, pancreatic surgeons, interventional radiologists) who counsel patients regarding the risks and benefits of each therapeutic option. (See "Endoscopic stenting for malignant biliary obstruction".)

EUS-guided biliary drainage with a single-step delivery system using an electrocautery enhanced, lumen-apposing metal stent (EC-LAMS) is associated with successful stent placement rates of over 90 percent in patients with malignant biliary obstruction [56,58,59]. The advantage of the EC-LAMS is that the delivery system is a one-step process, in comparison to the traditional EUS-guided approach which uses a multi-step process involving transduodenal needle access, injection of contrast into the bile duct, tract dilation, and then stent placement. While EUS-guided biliary drainage using EC-LAMS is a promising application of therapeutic EUS, larger studies with long-term outcome data are needed to compare its efficacy and safety with other approaches, such as percutaneous drainage. In a study of 46 patients with malignant distal bile duct obstruction who failed ERCP-guided drainage, 42 patients (91 percent) had adequate drainage (defined as at least a 50 percent decrease in bilirubin level within two weeks) following EUS-guided creation of a duodenobiliary fistula (or a gastrobiliary fistula in one patient) with placement of an EC-LAMS [56]. Major adverse events related to EUS-guided biliary stent placement included fatal hemorrhage (one patient) and stent occlusion or migration requiring reintervention (four patients).

Accessing the pancreatic duct — EUS-guided access of the pancreatic duct has been performed in patients with pancreatic stricture, with or without surgically-altered anatomy:

●Surgically-altered anatomy – For patients with an obstructed anastomosis following pancreaticojejunostomy, EUS-guided intervention to access and drain the pancreatic duct has been performed as an alternative to endoscopic retrograde pancreatography (ERP)-guided intervention. In a systematic review of 13 studies including 202 patients with anastomotic stricture following pancreaticojejunostomy, EUS-guided intervention was associated with higher rates of pancreatic duct cannulation (79 versus 26 percent) and pancreatic duct stent placement (72 versus 20 percent) compared with ERP-guided intervention [60]. In 111 patients who underwent EUS-guided pancreatic duct drainage, postprocedure abdominal pain occurred in 13 patients (12 percent), while comparison of adverse event rates between EUS- and ERP-guided interventions was not reported. These data have suggested that EUS-guided intervention is a promising option for managing obstructed pancreaticojejunal anastomosis, while its use may be limited by the availability of endoscopic expertise. (See "Surgical resection of lesions of the head of the pancreas", section on 'Pancreaticojejunal anastomotic stricture'.)

A specific approach for accessing an obstructed pancreaticojejunal anastomosis is an EUS-guided rendezvous procedure. After access to the main pancreatic duct has been established with EUS-guided fine needle aspiration (FNA), a guidewire is passed in an antegrade fashion across the pancreaticojejunal stricture into the afferent limb where it can then be retrieved with a pediatric colonoscope or enteroscope, followed by placement of a stent across the pancreaticojejunal anastomosis. Although EUS-guided rendezvous approach is technically easier and less invasive than EUS-guided placement of a transgastric stent, some strictures are too narrow to traverse by using a retrograde or antegrade-rendezvous approach.

●Normal anatomy – EUS-guided stent placement and drainage of the main pancreatic duct has been performed as an alternative to surgery in patients with pancreatic stricture [61]. In a study of 13 patients who failed ERCP and underwent EUS-guided creation of a pancreaticogastric fistula for stent placement in the main pancreatic duct, the drainage procedure was successful in 10 patients (77 percent). Complications included one case of bleeding treated with endoscopic clip placement and one case of a contained perforation.

EUS-guided fiducial placement for cyberknife therapy — CyberKnife stereotactic radiotherapy delivers multiple beams of precisely directed radiation to a tumor using real-time image guidance. Radiographic markers (fiducials) are implanted at the tumor site as reference points to assist in targeting the radiation beams. (See "Radiation therapy techniques in cancer treatment".)

Fiducials have traditionally been placed surgically or percutaneously under ultrasound or computed tomography guidance [62-64]. Case series have demonstrated the feasibility of placing them with EUS. In one series, for example, EUS was used to place fiducials in patients with tumors located in the mediastinum, retrocardiac region, retrocrural region, esophagogastric junction, porta hepatis, and pancreas [65]. Placement was successful in 11 of 13 patients (85 percent) with one failure secondary to gastric outlet obstruction precluding endosonographic visualization of the tumor, and the second failure due to intervening vasculature with increased risk for bleeding with needle puncture. Of 13 patients, seven were treated for pancreatic cancer with successful fiducial placement in 100 percent.

The technique for fiducial placement was similar to the previously mentioned approach for EUS-guided brachytherapy. After loading the fiducial into a 19-gauge needle and partially withdrawing the stylet to allow room, the fiducial is placed under EUS guidance near the treatment field (image 2 and image 3) to assist monitoring of respiratory variation under fluoroscopy (image 4).

A single complication occurred after fiducial placement in a patient with porta hepatis lymph node metastasis from colorectal cancer. The patient was admitted 25 days after the procedure for cholangitis and treated with antibiotics and percutaneous biliary drainage. Following this case, all patients were treated with prophylactic antibiotics followed by a three-day course after the procedure.

Another series described experience with EUS-guided fiducial placement in patients with unresectable gastrointestinal malignancies (ie, pancreatic cancer, cholangiocarcinoma, recurrent gastric and colon cancer) [66]. Fiducials were successfully placed in 15 of 16 (94 percent) patients with no adverse clinical outcomes.

In another study, 51 patients underwent EUS-guided fiducial placement for locally advanced or recurrent pancreatic cancer. Successful placement was achieved in 90 percent, with spontaneous migration in 7 percent and technical failure in 8 percent (all in patients with recurrent cancer after pancreaticoduodenectomy). Only one complication of mild pancreatitis occurred in a patient who underwent simultaneous celiac plexus neurolysis [67].

EUS-guided fiducial placement for stereotactic radiosurgery appears to be safe and feasible. However, additional studies are needed to determine the efficacy of this approach with CyberKnife radiotherapy. Infectious complications may be reduced by using sterile precautions and prophylactic antibiotic therapy.

EUS-guided fine-needle tattoo placement — Small pancreatic tumors may be difficult to locate during minimally invasive surgeries such as laparoscopic enucleation and laparoscopic distal pancreatectomy. EUS-guided fine-needle tattoo (EUS-FNT) was first described for the preoperative localization of a pancreatic insulinoma [68]. Since then, several case reports and case series have demonstrated the efficacy and safety of this technique for both pancreatic endocrine tumors and pancreatic adenocarcinoma [69-71]. Once the tumor is localized with EUS, fine-needle injection is performed with various dyes (eg, India ink, purified carbon particles [SPOT], methylene blue, or indocyanine green) to tattoo the tumor (picture 1). EUS-FNT may provide better localization of small tumors, thereby limiting the amount of resected pancreatic parenchyma, an important goal for a minimally invasive surgery.

EUS-guided angiography — EUS-guided angiography is a potential alternative to the traditional percutaneous route for accessing the vascular system. Studies in animals have demonstrated feasibility in accessing the portal vein, thoracic and abdominal aorta, celiac axis, superior mesenteric artery, splenic artery, splenic vein, and hepatic vein [72-74]. EUS-guided creation of an intrahepatic portosystemic shunt was technically feasible in a live porcine model and may become an alternative to transjugular intrahepatic portosystemic shunt placement [75]. However, the risks, benefits, and complications need to be carefully investigated before its role in humans becomes clear.

EUS-guided gastroenterostomy — Roux-en-Y gastric bypass can create a unique challenge for performance of endoscopic biliary procedures, requiring either surgical access through the excluded stomach or papillary access via deep small bowel enteroscopy, which are associated with logistic difficulty and low procedural success, respectively. (See "ERCP in patients with Roux-en-Y anatomy".)

Lumen-apposing metal stents (LAMS) have been developed, and EUS has been used to guide placement of LAMS to create gastrogastric fistulas. In some cases this permits access for endoscopic retrograde cholangiopancreatography (ERCP) at the time of the index procedure. In a series of five patients, 60 percent were able to undergo successful ERCP at the time of LAMS placement, with two patients requiring a second procedure [76]. The stents remained in place for at least three weeks, and the authors reported no adverse events or weight gain. Others have reported EUS-guided LAMS placement for creation of a gastroduodenostomy or gastrojejunostomy in the case of gastric or duodenal outlet obstruction [77,78]. It must be noted that these procedures were performed by or under the guidance of endoscopists who are very experienced in performance of both EUS and biliary procedures.

SUMMARY AND RECOMMENDATIONS

●Background – The use of therapeutic endoscopic ultrasound (EUS) is evolving. Clinical applications include drainage of pancreatic fluid collections and postoperative fluid collections, treatment of cystic neoplasms of the pancreas, EUS-guided cholangiopancreatography, localized therapy for pancreatic tumors, and treatment of gastric varices. (See 'Introduction' above.)

●EUS-guided drainage procedures – EUS is well suited to safely drain fluid collections of various types in areas accessible from the stomach, duodenum, or rectum. Most published data include studies on drainage of pancreatic fluid collections, but case series have described other drainage procedures (eg, abscesses, bilomas). (See 'Drainage procedures' above.)

●EUS with fine needle injection – EUS permits precise targeting for delivering various substances directly into pancreas, liver, or subepithelial lesions. More specifically, EUS-guided fine-needle injection of various agents has been reported for treatment of insulinomas, hepatic metastases, esophageal cancer, pancreatic cystic neoplasms, and pancreatic adenocarcinoma. (See 'Fine-needle injection' above.)

In addition, EUS-guided celiac plexus interventions have been used for patients with pain related to pancreatic disease. (See "Endoscopic ultrasound-guided celiac plexus interventions for pain related to pancreatic disease".)

●EUS-guided cholangiopancreatography – EUS-guided cholangiopancreatography is a technique for gaining access to the bile ducts and/or the pancreatic duct when endoscopic retrograde cholangiopancreatography was not possible due to altered surgical anatomy or obstructing duodenal or ampullary tumors. The left intrahepatic ducts, common bile duct, and main pancreatic duct are well visualized with EUS and can be accessed with either a transgastric or transduodenal approach. (See 'EUS-guided cholangiopancreatography' above.)