Prior to endoscopy, appropriate high resolution imaging (CT or Magnetic Resonance Imaging (MRI)) is typically obtained to delineate tumor burden and a reasonable roadmap of the biliary tree. Insertion of a biliary endoprosthesis is performed via endoscopic retrograde cholangiopancreatogram (ERCP) where selective cannulation of the bile duct achieved via a catheter and over a guidewire, a stent is typically advanced through a tight stricture, and deployed across the ampulla and into the duodenum. Biliary stent allows for drainage with subsequent relief of jaundice and pruritus with success rate of insertion at 90–95% in most experienced centers [20]. Endoscopic failures can occur and associated with limited access to major papilla from duodenal obstruction, inability to cannulate common bile duct or complex strictures. Complications include, but are not limited to, bleeding, cholangitis, pancreatitis and perforation.

Initially, plastic stents were used for palliation and rapidly gained popularity due to ease of use and relatively low cost. The median patency rates were approximately 3–4 months [21,22]. The main limiting factors to plastic stents not only included patency but also migration. In addition, the stent was a nidus for infection as intestinal bacterial flora had access to the biliary system, as well as obstruction due to a biofilm that eventually encases the stent lumen [23]. It was soon realized that larger diameter (then 10 French or 3.3 mm) would provide better stent patency [24,25].

With the evolution of self-expandable metal stents (SEMS) patency length was improved. Using small caliber catheter delivery system (7.5–10 French) accommodating the working channel of the endoscope, the SEMS is deployed across the malignant biliary stricture up to 30 French. The larger diameter allows for enhanced patency, up to 10 months. A meta-analysis comparing plastic to metal biliary stents demonstrated no statistical difference in technical success, therapeutic success, complications, or 30 day mortality but did show a lower risk of biliary obstruction for the latter [26]. Cost effectiveness in this analysis was inconclusive, thought to be due to patient length of survival. More recently, a study from Korea compared costs of metal versus plastic stent for unresectable malignant biliary obstruction, where costs of ERCP are lower than that of metal stents. Stent occlusion was higher in the plastic stent group in a shorter amount of time (133 vs. 278 days; p = 0.0004) as well as a higher rate of cholangitis and associated length of hospitalization, with no statistically significant cost difference ($1488.77 metal stent group vs. $1319.26 plastic stent group; p = 0.422) [27].

Interestingly and supported by the previous studies, survival was no different and fuels the debate about what stent is most appropriate. Most endoscopists would agree that prognosis plays a large part in this decision, as metal stents provide the most benefit in patients expected to survive at least 4–6 months [28]. The rationale for this approach is that if the patient will not survive for the next 4–6 months, the patency advantage will not be of benefit as plastic stents can be patent for 3–4 months [29-31].

SEMS initially were uncovered and this feature did not make them removable in case of malfunction and allowed for tumor ingrowth and mucosal hyperplasia [32]. These factors limited the patency and lead to the evolution of first partially covered and then fully covered SEMS, with the goal of negating tumor ingrowth. A multicenter randomized control trial comparing covered and uncovered SEMS showed better patency and lower occlusion for covered SEMS (mean 304 days vs. 166 days for uncovered stent) [33]. However more recent studies have failed to confirm this data [34,35]. Most recently, Telford et al. compared uncovered SEMS to partially covered SEMS (Wallstent, Boston Scientific, Natick MA) in patients with inoperable distal malignant obstruction and found no difference in rate of biliary obstruction (p = 0.53) or patient survival (p = 997), although the partially covered stent was prone to migration (p = 0061) [34]. It is still unclear which stent is ideal as limitations exist for both uncovered (described above) and for covered stents such as migration, and cholecystitis (by occluding cystic duct drainage-noted by both stent types) [36,37]. One new trend currently noticed in the literature is the increase use of metal stents in the setting of neo-adjuvant protocols (with delayed surgery) or as a first-intention intervention regardless of resectability as demonstrated by our group [36,38]. All data available so far seem to demonstrate the cost-effectiveness of such a concept [39]. Furthermore, a recent study by van der Gaag et al. argues against preoperative biliary decompression using plastic stents [40].

In order to place a stent for relieving an obstruction, the location and the extent of tumor involvement is defined via imaging (i.e., oral contrast study).

Endoscopic method gives better access to the duodenum with the ability to visualize to the site of obstruction. There are ongoing studies about the use of increasing types, shapes and sizes of stents with different compositions, however there are only two that are FDA approved: Enteral Wallstent (Boston Scientific, Natick, Massachusetts) and its newer version the Enteral Wallflex (Boston Scientific) for duodenal stenting that can be advanced into the working channel of a standard endoscope [44]. Once the site is reached, the selected stent should be 3–4 cm longer than the stricture, and may even require that stents be overlapped for adequate relief of obstruction [45]. Contrast and fluoroscopy is used to assure passage of obstruction as well as appropriate positioning of the stent. Common problems encountered with stenting of the duodenum include biliary obstruction in up to 40% of these patients requiring percutaneous or endoscopic drainage [46]. One proposal to avoiding this complication is to place a biliary stent before attempting enteral stenting. Failures are related to inability to cross the stricture due to complete obstruction, inaccessible distal strictures, functional gastric outlet obstruction, or peritoneal carcinomatosis or gastroparesis [45]. In these cases, the best option for palliation is to place a venting gastrostomy. As with endoscopic procedures, duodenal stenting carries risks of bowel perforation, bleeding, stent malposition and/or migration and worse case, fistula formation [45].

The success of duodenal stenting are measured by the patient returning to oral intake, with an average rate of 87–100% with minimal complication rates as low as 1.2% [47]. Philips et al. have one of the largest series of patients with enteral stenting involving 46 patients with reported 100% success rates of SEMS deployment and 91% clinical success rate [48]. Even though surgical bypass is historically a standard of treatment of GOO, recent data supports endoscopic duodenal stenting as a preferred alternative [9]. Stenting proved to be 1/3 of the cost of surgery with decreased hospital admission days [40]. Another major advantage is that patients are able to eat as early as two days post-operatively day as reported by Fiori et al., which is a major goal of palliation [49].

Celiac plexus block (CPB) was first described in 1914 by Kappis et al. and has since evolved to lead to development of celiac plexus neurolysis (CPN). Both methods are now used via endoscopic ultrasound (EUS) guidance for different purposes; where CPB is mostly used for chronic pancreatitis and CPN for pancreatic cancer pain [57]. The difference between CPB and CPN is in the injected agents and the location of injection. These procedures require that the endoscopist identify the antecrural space, noted to be anterior to the celiac plexus and inject the preferred agents in the appropriate location [58]. Once the appropriate landmarks are endoscopically identified, CPB involves injecting an anesthetic such as bupivacaine and a steroid in the area of the celiac plexus or on top of the ganglia [57]. CPN evolved with the increased recognition of the celiac ganglia via EUS. Levy et al. described the efficacy of direct injection into the ganglia [59]. This technique involves the same identification of structures; however, it requires directly injecting the celiac ganglia with a sclerosant, such as phenol or dehydrated alcohol with an anesthetic.

Multiple studies conclude that both CPB and CPN have a significant palliative role in chronic pancreatic abdominal pain. However, it is important to note that they do not eliminate the use of pre-procedure narcotic pain management. CPB appears to be most effective for chronic pancreatitis pain rather than pancreatic cancer pain as described by a recent meta-analysis [57]. The study concluded that there was an average of 51% reduction in abdominal pain. CPN on the other hand was more efficacious in pancreatic cancer pain, with an average of 73% reduction of pain. Levy et al. described EUS-guided CPN on patients with unresectable pancreatic carcinoma and chronic pancreatitis (n = 33). Ninety four percent (16/17) of patients with cancer and 4/5 with chronic pancreatitis reported pain relief. Interestingly, 34% had initial pain exacerbation thought to be related to improved therapeutic response (p < 0.05) [59]. A few studies reported that the efficacy of this procedure was dependent on various factors such as timing and location of the tumor. Mercadente et al. evaluated the use of narcotics post CPN and reported a transient reduction in supplemental narcotic pain management for up to four weeks post procedure then rebounded back to their original pain level if higher doses of opioids were note used. The authors concluded that CPN should be initiated earlier in the course of the pain to prevent the rebound, while the anatomy of the pancreas is still in tact to identify appropriate structures [60].

Catalano et al. reported the importance of tumor location in the efficacy of CPN, where patients with malignancy at the body or tail of the pancreas had better pain relief than those with pancreatic head cancer [61]. EUS-guided CPN carries well studied risks as described by Gunaratnam et al., with hyperalgesia in 9%, postural hypotension in 20% and diarrhea in 7% of the patients, all of which were noted to be transient symptoms [62]. The benefit includes sustained pain relief for up to 24 weeks independent of adjuvant therapy [62].