EUS image of celiac artery ( arrow) and SMA ( arrowhead) taking off from the aorta. (Courtesy Dr. Linda Lee, Brigham and Women’s Hospital, Boston, MA)
EUS-Guided Direct Celiac Ganglion Neurolysis
Recently, it has been recognized that the individual celiac ganglia can be visualized and accessed by EUS allowing for direct injection into the individual celiac ganglia to perform celiac ganglion neurolysis (CGN). The celiac ganglia are typically oval or almond shaped ranging in size between 2–20 mm and most readily detected to the left of the celiac artery, anterior to the aorta. Compared to the surrounding retroperitoneal fat, the ganglia are echo poor and often display similar echogenicity to the left adrenal gland. Within the ganglia, often central echo-rich strands and foci are present and the margins of the ganglia are irregular. Color Doppler demonstrates little to no flow within these structures. Ganglia are detected by EUS in between 81 and 89 % of patients . Our approach is to perform CGN rather than CPN if ganglia are visualized although further data are needed on this approach.
What is the Technique for EUS-Guided CGN?
All aspects of the procedure including patient candidacy, sedation , antibiotic use, and follow-up are the same as standard CPN/CPB. The technique for CGN and volume of solution injected has not been standardized. Our approach is to target as many ganglia as possible by injecting a total of 10–20 ml of premixed alcohol and bupivacaine (mixture as outlined above) among all the ganglia in amounts relative to their size (Fig. 33.2, see Video 30.1). For example, if three ganglia are visualized (small, medium, and large), we would typically inject 5 ml in the largest ganglion, 3 ml in the medium sized ganglion, and 2 ml in the small ganglion. For larger ganglia, we typically advance the needle tip into the deepest point within the ganglia and then inject while slowly withdrawing the needle, creating an even distribution of injectate throughout the ganglion. For smaller ganglia, we usually target the ganglia’s center. During injection, a clear “ballooning” of the ganglia should be visualized; otherwise, needle placement is considered suboptimal.
EUS-guided direct injection of alcohol and bupivacaine solution into two celiac ganglia in a patient with pancreatic cancer
Clinical Trial Data
The clinical trial data for CGN and CPN for patients with pancreatic cancer are summarized in Table 33.1. Within the literature, there is great variability among the studies in terms of injection technique, type of injectate and volume, definition of pain relief, and follow-up. Most studies are small retrospective studies with short follow-up.
Clinical trial data for EUS-guided celiac plexus neurolysis
Pain relief (% of patients)
Doi et al. 2013 
Ganglia vs. plexus
1–2 ml bupivacaine 0.25–0.5 % 10–20 ml alcohol
73.5 vs. 45.5 % partial 50 vs. 18.2 % Complete
Hypotension 2.9 % vs. 6.0 UGI bleed 2.9 vs. 0 % Increased pain 29.4 vs. 21.2 %
Seicean et al. 2013 
10 ml bupivacaine 1 % 10–15 ml alcohol
LeBlanc et al. 2011 
Plexus 1 vs. 2 injections
20 ml bupivacaine 0.75 % 10 ml alcohol 98 %
69 vs. 81 %
Hypotension 2 % Increased pain 33 %
Iwata et al. 2011 
2–3 ml bupivacaine < 20 ml alcohol
Hypotension 17 % Diarrhea 23 % Transient inebriation 9 %
Ascunce et al. 
Ganglia vs. plexus (bilateral)
10 ml lidocaine 1 % 20 ml alcohol 98 %
65 % vs. 25 %
Hypotension 2 % Increased pain 2 % Diarrhea 23 %
Sakamoto et al. 2010 
Plexus (bilateral) vs. broad plexus (around SMA)
3 ml lidocaine 1 %; 9 ml alcohol; 1 ml contrast
50 vs. 76 %
No major complications
Sahai et al. 2009 
Plexus (central) vs. plexus (bilateral)
10 cc bupivacaine 0.5 % 20 cc alcohol
50.7 vs. 77.5 %
Retroperitoneal bleed 0.7 %
Levy et al. 2008 
8 ml Bupivacaine 0.25 % 12 ml alcohol 99 %
Hypotension 35 % Increased pain 41 %
Gunaratnam et al. 2001 
6–12 ml bupivacaine 0.25 % 20 ml alcohol 98 %
Hypotension 20 % Increased pain 9 % Diarrhea 17 %
For celiac plexus neurolysis , partial pain relief has been reported between 50 and 78 % within the first 4 weeks [11, 15, 17–19]. A meta-analysis including 119 patients found that EUS-CPN alleviated abdominal pain in 73 % of patients.  In a randomized trial, 96 patients with inoperable pancreatic cancer were randomized into conventional pain management or EUS-CPN. At 3 months, patients treated with CPN had greater pain relief with a trend toward lower morphine consumption, although no difference was observed in quality of life .
Results of celiac plexus block are less successful than CPN with two meta-analyses reporting 51 and 60 % pain relief [10, 20]. Of those who respond, duration of relief is short-lived at about 4 weeks . A randomized study comparing percutaneous fluoroscopy directed CPB to EUS-CPB in chronic pancreatitis found that significantly more patients reported improvement in pain scores in the EUS group (70 versus 30 %, p = 0.04) . Unfortunately, by 24 weeks following CPB, nearly all patients had returned to their baseline pre-procedure pain score. Another prospective non-randomized study confirmed this lack of durability of pain relief with only 10 % reporting some pain relief at 24 weeks .
For celiac ganglion neurolysis, partial pain relief has been reported between 65 and 94 % [24–26]. In the only prospective trial to date comparing CGN to CPN, 68 patients with upper abdominal cancer (over 85 % were pancreatic cancer) were randomly assigned to treatment using either EUS-CGN or EUS-CPN with one midline injection. The positive response rate was significantly higher in the EUS-CGN group (73.5 %) than in the EUS-CPN group (45.5 %). The complete response rate was also significantly higher in the EUS-CGN group (50 %) than in the EUS-CPN group (18.2 %) although the response rate for EUS-CPN was much lower than reported in other studies. There was no difference in adverse events or duration of pain relief between the groups . Follow-up was only 7 days, and much longer term follow-up studies are needed as well as comparison with bilateral EUS-CPN injections.
What Common and Serious Complications Can Occur?
Most complications related to CPN/CPB and CGN are transient, and serious complications are rare. A large series of 220 patients undergoing EUS-CPN/CPB had an overall complication rate of 1.8 % with complications defined as any side effect requiring management beyond standard post-procedure observation . The most common side effects reported are transient hypotension (up to 35 %), diarrhea (up to 20 %), and transient exacerbation of pain following procedure which are consistent with rates seen with the PQ approach . Hypotension and diarrhea are related to sympathetic blockade and the relative unopposed visceral parasympathetic activity. Hypotension generally responds to intravenous fluid administration. The diarrhea related to this procedure is usually self-limited and resolves in less than 48 h. CPN via a PQ approach has been associated with a 2 % rate of serious complications including neurologic complications (lower extremity weakness, paresthesia, paralysis), pain (pleuritic chest, shoulder), pneumothorax, and hiccupping . A very small number of serious complications (≤ 0.6 %) including fatalities and paralysis mainly with alcohol injection have been reported with the EUS approach in case report and abstract form [28, 29]. Serious infections including retroperitoneal abscess and empyema have occurred as well as severe ischemic damage to abdominal organs.
In the case of Mr. H, celiac neurolysis appears to be an ideal choice for pain relief. His cancer-related pain is clearly not optimized with oral narcotic agents, which are causing side effects of constipation and drowsiness. To date no trials have compared percutaneous to EUS-guided celiac neurolysis; however, our preferred approach in this patient would be EUS-guided celiac ganglion neurolysis although celiac plexus neurolysis is also reasonable. Direct visualization of the celiac ganglia can only be achieved using EUS at the current time and as a result, a non-EUS approach is limited to CPN. Additional well-designed studies are needed to further compare CGN and CPN and to determine the optimal method and timing of celiac neurolysis/block, composition of the injectate, impact on quality of life, and the benefit of this technique in chronic pancreatitis patients.
EUS-Guided Alcohol Ablation of Pancreatic Cysts
Mrs. J is a 75-year-old female with severe chronic obstructive lung disease who recently presented to the hospital with dyspnea and underwent a high-resolution CT scan of the chest which incidentally discovered a 3.5-cm unilocular head of pancreas cyst . The patient’s sister passed away from pancreatic cancer in her sixties, and she is deeply worried about this lesion in her pancreas. The patient undergoes pancreatic EUS, which shows a single 3.5-cm unilocular lesion in the head/uncinate region with no apparent communication to a non-dilated main pancreatic duct. EUS-guided cyst fluid aspiration reveals mucinous fluid with string sign of 12 mm and CEA 350 ng/ml. The patient’s general gastroenterologist diagnoses her with probable branch duct intraductal papillary mucinous neoplasm (IPMN) and recommends that she undergo Whipple procedure to remove this lesion. The patient asks whether there are any other options for treatment of this lesion.
The widespread use of cross-sectional imaging has dramatically increased the number of incidentally noted pancreatic cystic lesions (Chap. 31). Recent magnetic resonance imaging (MRI) and computerized tomography (CT) studies indicate a prevalence of pancreatic cysts in up to 14 % [30, 31]. Pancreatic cysts encompass a wide spectrum of histopathology and the epithelium that outlines pancreatic cystic neoplasms (PCNs) may have negligible malignant potential (serous cystadenomas) or represent premalignant lesions such as intraductal papillary mucinous neoplasms (IPMNs) or mucinous cystic neoplasm (MCNs) [32, 33].
Consensus guidelines and expert opinion recommend surgical resection of MCNs and IPMNs that are symptomatic, larger than 3 cm in diameter, contain mural nodules, or involve the main pancreatic duct (main duct IPMNs) [34, 35]. However, because of the operative risk associated with pancreatic resection there has been increasing interest to develop a minimally invasive technique to treat these lesions. Pancreatic cyst ablation may be an attractive option for patients with multiple comorbidities deemed high-risk for surgery. Potential benefits include decreasing the malignant potential of the cyst, lower cost over time due to reduced cyst surveillance , and the psychological benefit to the patient. The following will review the technique, clinical trial data, and controversies surrounding pancreatic cyst ablation.
What is the Technique of EUS-Guided Pancreatic Cyst Ablation?
All studies to date describing EUS-guided pancreatic cyst ablation have used a 22-gauge needle [36–41]. With a curvilinear-array echoendoscope, the cyst is typically punctured via a transgastric or transduodenal route. Using a syringe, cyst fluid is aspirated with the goal of near complete evacuation of the cyst. Removing the cyst fluid before injection increases the surface area directly exposed to the ablative agent and improves the efficacy of ablation. The amount of aspirated fluid, viscosity, and color should be noted. To ensure the cyst does not completely collapse, the needle remains within the cyst before injection of the ablative agent. With viscous cyst fluid, it may not be possible to evacuate the cyst contents as much as desired; therefore, saline may be injected into the cyst to decrease viscosity or expand a small cyst to confirm needle placement .
With the needle in the nearly collapsed cyst, ethanol is injected into the cyst using a volume equal to that initially aspirated from the cyst. During the procedure, the needle tip is carefully maintained within the cyst to avoid parenchymal injury or leak in the cyst wall. Studies to date have performed lavage for 5 min, alternately filling and emptying the cavity during that time. For cysts with viscous fluid, this is performed as 3–4 lavages over the 5-min period. When the cyst fluid is thin, 7–8 lavages are performed over the same period. At the completion of alcohol lavage, the cyst cavity should be completely drained of fluid as much as possible. If a chemotherapeutic agent such as paclitaxel is used after alcohol lavage, as much alcohol is removed prior to paclitaxel injection. With alcohol or paclitaxel injection, the total injection volume should not exceed the volume of aspirated fluid and hence the cyst should not be expanded beyond its original diameter. When paclitaxel is injected after alcohol lavage, it is left in place and not drained. The needle is then removed from the cyst cavity [42, 43].
What Ablative Agents are Available for Pancreatic Cyst Ablation?
Ethanol (80–98 %) is an inexpensive, widely available, low-viscosity agent that is easy to inject through a small gauge needle. It is hypothesized to induce cell death by membrane lysis, protein denaturation, and vascular occlusion  and has been used for the destruction of solid and cystic tumors in a variety of organs . The only other agent used to date for pancreatic cyst ablation has been paclitaxel, which is a hydrophobic, viscous chemotherapeutic agent that inhibits cell processes that depend on microtubule turnover. Its viscosity enables it to exert a durable effect on the epithelium within the cyst cavity with a low risk of leakage .
Which Cysts are Considered Candidates for Ablation?
The size, morphologic characteristics, and suspected histologic type of the cyst guide the approach to cyst injection and ablation therapy (Table 33.2). In published studies the cysts that have been treated include suspected mucinous cysts measuring between 1 and 6 cm in maximal diameter . The ideal size for cyst ablation is based on two competing factors: the risk of malignancy and the success of ablation. Cysts larger than 3 or 4 cm are typically at higher risk for malignancy; however, cysts containing cancer are inappropriate for ablation. On the other hand, a cyst size of at least 2 cm is often needed to ensure feasibility and safety of the ablation procedure. As a result, the ideal cyst size for ablation is likely between 2 and 4 cm .
Characteristics of candidate cyst and patient for ablation
The ideal cyst for ablation
The ideal patient for ablation
A benign appearance without any malignant features
Patient who is high risk for surgery
A diameter between 2 and 4 cm
Without the presence of: ongoing pancreatitis, ascites, portal hypertension, coagulopathy
Unilocular or oligolocular morphology
No communication with main pancreatic duct
Cyst ablation has the greatest chance of success in unilocular or oligolocular cysts with fewer than 2–3 locules. In the presence of three or more locules, a single needle pass may not provide sufficient drug delivery to all locules within a cyst. It is important to determine the optimal angle at which the needle can be introduced into the maximal number of targeted locules. A second needle puncture may be considered when it can be performed without increasing the risk for adverse events .
The presence of a communication between the cyst and the main pancreatic duct may result in flow of the injected ablative agent through the communicating duct into the main pancreatic duct. This outflow may diminish the ablative effect and also increase risk of ductal change. Oh and colleagues  performed endoscopic retrograde cholangiopancreatography (ERCP) in all patients to exclude any cyst communication with the main pancreatic duct. Other studies did not perform ERCP before ablation [36, 38]. A practical point to note is if a cyst does not restore to its original size during ethanol injection, vigorous lavage with repeated injection and aspiration should be avoided because of a probable communication with the main pancreatic duct .
Which Patients Should be Selected for Ablation?
There is no clear consensus on whom should undergo pancreatic cyst ablation. Patients with high-risk, symptomatic, or benign cysts who refuse or are not fit for surgery may be considered for ablation (Table 33.2). Patients with imaging consistent with MD-IPMN such as a dilated main pancreatic duct or suggesting malignancy such as mass-like lesions, suspicious liver, or pulmonary lesions or enlarged lymph nodes should not be offered cyst ablation. Similarly, patients with active or ongoing pancreatitis, ascites, portal hypertension, or coagulopathy have generally been excluded from cyst ablation.
Clinical Trial Data
To date, six published prospective studies have evaluated the role of EUS-guided pancreas cyst ablation. Table 33.3 summarizes the clinical trial data of cyst ablation in the literature currently. Endpoints in all studies include EUS or radiologic assessment of changes in baseline cyst size after ablation. For ethanol ablation alone, cyst resolution (defined as no visible residual cyst) ranged from 33 to 38 % [36, 38, 41] based on cross-sectional imaging. The addition of paclitaxel appears to increase CT-defined cyst resolution (defined as size < 5 % of original cyst volume) with range between 60 and 79 % [37, 40, 46].
Clinical trial data for pancreatic cyst ablation
No of patients
Complete resolution on imaging
Median months follow -up (range)
Gan et al. 2005 
5–80 % ethanol
35 % (8/23)
At least 3 months, not stated
Oh et al. 2008 
80/99 % ethanol with paclitaxel
79 % (11/14)
Pancreatitis (10 %)
Oh et al. 2009 
99 % ethanol with paclitaxel
60 % (6/10)
Pancreatitis (7 %) Abdominal pain
DeWitt et al. 2009 
80 % ethanol
33 % (12/36)
Pancreatitis (4.5 %) Intracystic bleeding (2 %) Abdominal pain at 2h (14 %) Abdominal pain 7d (20 %)
At least 3 months, not stated
Oh et al. 2011 
99 % ethanol with paclitaxel
62 % (29/47)
Fever (2 %) Pancreatitis (2 %) Abdominal pain (2 %)
DiMaio et al. 2011 
80 % ethanol
38 % (5/13)
Abdominal pain (8 %)
13 month after first lavage
EUS-guided cyst ablation has generally been well tolerated with a low rate of adverse events. The initial pilot study that evaluated the safety of injecting increasing concentrations up to 80 % ethanol found no treatment-related complications . Subsequent studies have shown pancreatitis rates between 2 and 10 %, abdominal pain in 2 and 20 %, fever in 2 % and intracystic bleeding in 2 %. To date, no cases of severe pancreatitis, bleeding requiring transfusion, or deaths have been reported. Prophylactic antibiotics should be administered.
After discussion with the patient, it was determined that she was a good candidate for cyst ablation with a unilocular cyst measuring 3.5 cm not communicating with the pancreatic duct and no evidence of malignancy. She was also at higher risk for Whipple surgery due to her lung disease . She agreed to try EUS-guided pancreatic cyst ablation, which was performed using monitored anesthesia care and a 22-gauge needle with a single pass into the cyst after administering intravenous ciprofloxacin. Following complete aspiration of the cyst fluid, equal volume of 80 % ethanol was injected and the cyst lavaged over 5 min. She did well post-procedure with no complications. Follow-up CT scan 3 months later demonstrated decrease in size of the cyst by about half.
Pancreatic cysts are being encountered frequently in clinical practice, including in the elderly and patients at high risk for surgical resection. Therefore, a non-surgical treatment alternative is desirable. EUS-guided cyst ablation is an emerging modality that may present an alternative to surgery, especially if complete ablation can be achieved in the vast majority of patients. At the present time, this is an investigational procedure and studies to date have shown that cyst ablation is relatively safe with cyst resolution in up to 67 % of patients. It is uncertain whether incomplete ablation of the neoplastic cyst lining will reduce cancer risk or whether partially treated cysts will become more difficult to monitor. Future research is now needed to focus on refinement of the technique , choice, and number of ablative agents, selection criteria of appropriate cysts for treatment, and the long-term outcomes of this treatment.
Endoscopic Ultrasound-Guided Radiofrequency Ablation
Image-guided radiofrequency ablation (RFA) is a well-recognized minimally invasive treatment modality in oncology, one that utilizes the generation of high-frequency electrical alternating current through target tissue to induce ion agitation and tissue friction ultimately leading to thermal injury and consequent coagulative necrosis. Effective ablation is achieved by optimizing heat production and minimizing heat loss with the objective of generating a clear tumor ablation margin while reducing potential side effects. The availability, safety, efficacy, and low cost of percutaneous RFA have facilitated its common utilization in conjunction with ultrasound, CT, or MRI guidance for the management of a variety of solid tumors, most commonly hepatocellular carcinoma, renal cell carcinoma, non-small-cell lung cancer, and osteoid osteoma.
RFA has also been used to treat pancreatic cancer during exploratory laparotomy or laparoscopy; a recent systemic review identified five studies including 158 patients with a median survival after RFA of 3–33 months, 0–19 % mortality, 10–43 % overall morbidity, and 4–37 % RFA-related morbidity, much of which was related to collateral injury to adjacent tissues . Given its minimally invasive nature and superior imaging capabilities of the pancreas, endoscopic ultrasound potentially provides an ideal vehicle for delivering RFA to pancreatic cancer as well as other percutaneously inaccessible tumors.
Eight studies, six utilizing porcine models, of EUS-guided RFA have been conducted to date (Table 33.4). Using a modified 19-gauge needle electrode connected to a monopolar radiofrequency generator, Goldberg et al in 1999 first demonstrated the feasibility of EUS-guided RFA of the pancreas in 13 pigs . The maximum diameter of the ablated area was 10–15 mm by EUS and 12 mm by histology. Correlation between EUS or CT and gross pathologic findings for size of the ablated region was excellent for all areas larger than 5 mm; size of the ablated zone at pathologic examination was within 2 mm of that visualized on imaging. Complications included three transmural gastric wall burns, an intestinal serosal burn, and an asymptomatic pancreatic fluid collection.
Summary of animal and human studies evaluating EUS-guided radiofrequency ablation
Maximum diameter/area of ablated area at EUS
Maximum diameter/area of ablated area at histology
Monopolar 19-guage needle electrode
Transmural gastric wall burns (n = 3); Intestinal serosal burn (n = 1); Asymptomatic pancreatic fluid collection (n = 1)
Variable; up to 14 days
Hybrid bipolar RFA-carbon dioxide cryoprobe
Full access? Get Clinical TreeGet Clinical Tree app for offline access