Endoscopic ultrasound

According to current guidelines, endoscopic ultrasound (EUS) in pancreatic malignancy should ideally be performed before biliary stenting to facilitate staging and tissue acquisition. Previously placed biliary stents may influence accuracy and technical difficulty of EUS-guided fine-needle aspiration (FNA). Ranney et al. retrospectively analyzed a prospective database looking for answers to this common logistical problem in clinical practice [1]. Data from consecutive patients with solid pancreatic masses who underwent EUS–FNA during a 5year period at a tertiary unit were analyzed. All procedures were performed by a single endosonographer with on-site cytopathology. Patients were subdivided into those with (n=150) and those without (n=64) biliary stents at the time of EUS.The former were further divided for comparison into those with plastic stents (n=105) and those with self-expandable metal stents (SEMS) (n=45). Overall, no significant differences in diagnostic accuracy between patients with and without biliary stents (95.3% vs. 93.7%), and patients with plastic vs. SEMS (95.2% vs. 95.5%) were found. The timing of stenting did not influence the diagnostic rate of malignancy either, as patients with stents placed immediately before EUS and those with previously placed stents had similar outcomes. Furthermore, there was no significant difference in the median number of passes required to establish a diagnosis between patients with and without biliary stents and between patients with SEMS and those with plastic stents. EUS–FNA produced false-negative results in 11 patients, which was significantly influenced only by lesion size and number of passes. The retrospective design of the study precludes evaluation of factors such as the presence of chronic pancreatitis or stent characteristics. Results may be biased by the high rate of malignancy in the patients with SEMS, a subgroupwith a very high pre-test probability for cancer. All procedures were performed by one operator at a referral center and therefore these results might not be generalizable. In the only similar published study, the timing of stent placement significantly influenced cytology results [2]. Finally, although tissue diagnosis is the main objective of EUS in patients with pancreatic cancer, staging is of capital importance for thosewith apparently resectable disease on computed tomography. Prior studies have shown that it is more difficult to accurately stage pancreatic cancer with a biliary stent in place. Despite these limitations, Ranney et al. provide valuable information on a common clinical dilemma. According to their results, biliary decompression by means of plastic or SEMSs should not be delayed in symptomatic patients until EUS is available, given that the stent, irrespective of type, will not hinder tissue acquisition and diagnosis.


Endoscopy essentials 300
Impact of biliary stents on EUS-guided FNA of pancreatic mass lesions (Ranney et al., Gastrointest Endosc 2012 [1]) ! According to current guidelines, endoscopic ultrasound (EUS) in pancreatic malignancy should ideally be performed before biliary stenting to facilitate staging and tissue acquisition. Previously placed biliary stents may influence accuracy and technical difficulty of EUS-guided fine-needle aspiration (FNA). Ranney et al. retrospectively analyzed a prospective database looking for answers to this common logistical problem in clinical practice [1]. Data from consecutive patients with solid pancreatic masses who underwent EUS -FNA during a 5year period at a tertiary unit were analyzed. All procedures were performed by a single endosonographer with on-site cytopathology. Patients were subdivided into those with (n = 150) and those without (n = 64) biliary stents at the time of EUS. The former were further divided for comparison into those with plastic stents (n = 105) and those with self-expandable metal stents (SEMS) (n = 45). Overall, no significant differences in diagnostic accuracy between patients with and without biliary stents (95.3 % vs. 93.7 %), and patients with plastic vs. SEMS (95.2 % vs. 95.5 %) were found. The timing of stenting did not influence the diagnostic rate of malignancy either, as patients with stents placed immediately before EUS and those with previously placed stents had similar outcomes. Furthermore, there was no significant difference in the median number of passes required to establish a diagnosis between patients with and without biliary stents and between patients with SEMS and those with plastic stents. EUS -FNA produced false-negative results in 11 patients, which was significantly influenced only by lesion size and number of passes.
The retrospective design of the study precludes evaluation of factors such as the presence of chronic pancreatitis or stent characteristics. Results may be biased by the high rate of malignancy in the patients with SEMS, a subgroup with a very high pre-test probability for cancer. All procedures were performed by one operator at a referral center and therefore these results might not be generalizable. In the only similar published study, the timing of stent placement significantly influenced cytology results [2]. Finally, although tissue diagnosis is the main objective of EUS in patients with pancreatic cancer, staging is of capital importance for those with apparently resectable disease on computed tomography. Prior studies have shown that it is more difficult to accurately stage pancreatic cancer with a biliary stent in place. Despite these limitations, Ranney et al. provide valuable information on a common clinical dilemma. According to their results, biliary decompression by means of plastic or SEMSs should not be delayed in symptomatic patients until EUS is available, given that the stent, irrespective of type, will not hinder tissue acquisition and diagnosis.
Endoscopic ultrasound-guided vascular therapy: is it safe and effective? (Gonzalez JM et al., Endoscopy 2012 [3]) ! Gonzalez et al. reported EUS-guided therapy of refractory gastrointestinal bleeding in eight patients, secondary to arterial bleeding (n = 5) or segmental portal hypertension (n = 3) [3]. A variety of bleeding sources were treated, including gastric varices, Dieulafoy lesion, and gastroduodenal artery bleeding. Cyanoacrylate was used as sclerosant in six patients and polidocanol in two. Seven patients had active bleeding at the time of EUS therapy and six had experienced previous episodes of bleeding. Overall, the initial success rate of EUS-guided therapy was 87.5 %. The only patient who suffered rebleeding underwent a successful second EUS-guided therapeutic procedure. No complications occurred. Interestingly, seven patients showed immediate disappearance of Doppler signal with EUS after treatment. The only patient who did not show disappearance of Doppler signal suffered bleeding recurrence, suggesting a possible prognostic role for EUS. These results highlight the efficacy of the EUS-guided approach in patients with severe gastrointestinal bleeding after failure of conventional endoscopic therapy. The advantages of EUS in this setting include accurate identification of the source of bleeding and precise delivery of the thrombotic agent into the target vessel. The findings of Gonzalez et al. are in keeping with prior case series. Levy et al. reported the result of EUS-guided therapy in five patients with refractory and/or relapsing bleeding from gastrointestinal stromal tumors, Dieulafoy lesion, and a superior mesenteric artery pseudoaneurysm [4]. A 100 % success rate was achieved without complications. More recently, Binmoeller et al. reported results of EUS-guided coil deployment followed by cyanoacrylate injection in a series of 30 patients with bleeding gastric varices. The procedure was successful in all. Bleeding recurred in four patients. Complete obliteration of gastric varices was achieved in 23 after a mean follow-up of 193 days [5]. Romero-Castro et al. included 35 patients with refractory bleeding from gastric varices treated either by EUS-guided coil deployment or by cyanoacrylate injection [6]. Varix obliteration was achieved in all patients, at the expense of a 26 % complication rate, largely accounted for by cyanoacrylate injection. Limitations to EUS-guided vascular therapy include impairment of endoscopic or sonographic field of view by intraluminal clots or retained contents, the potential to inadvertently induce bleeding from an extraluminal site not accesible to conventional endoscopy, and possible echoendoscope damage during cyanoacrylate injection. Despite these drawbacks, these small series show the therapeutic potential of EUS for refractory bleeding in select cases, even from extraintestinal sources. Pending future investigations, the currently available evidence warrants a proactive role by endosonographers to help disseminate knowledge about this novel approach among clinicians caring for patients.
Endoscopic ultrasound-guided transmural and percutaneous transhepatic gallbladder drainage are comparable for acute cholecystitis (Jang et al., Gastroenterology 2012 [7]) ! Percutaneous cholecystostomy is the treatment of choice for patients with poor surgical performance. EUS-guided transmural gallbladder drainage (EUS -GBD) by means of nasobiliary tubes or internal stents (plastic or SEMS) may be a similarly effective therapeutic option, potentially eliminating the drawbacks associated with percutaneous access and drainage. Jang et al. performed a randomized, prospective controlled trial to compare EUS -GBD and percutaneous transhepatic gallbladder drainage (PTGBD) [7]. A total of 59 patients with acute cholecystitis refractory to medical treatment and unsuitable for emergency cholecystectomy were included. They were randomized to EUS -GBD with 5 Fr nasobiliary drainage tubes (n = 30) or PTGBD with 8.5 Fr pig-tail drainage catheters (n = 29). Laparoscopic cholecystectomy was attempted and scheduled prior to discharge for all patients who experienced technical and clinical success. Technical feasibility, efficacy, and safety of both techniques were compared. No significant differences in technical (97 % vs. 97 %) and clinical (100 % vs. 96 %) success rates and in complication rates (7 % vs. 3 %) were found between the EUS -GBD and PTGBD procedures. The median duration from drainage to cholecystectomy (5 vs. 6 days) and the conversion rate to open cholecystectomy at subsequent laparoscopic cholecystectomy was no different between the groups (9 % vs. 12 %). However, EUS -GBD patients had significantly lower post-procedure pain scores (P < 0.001). This well-designed trial provides compelling evidence that EUS -GBD is a good alternative to PTGBD for high-risk patients with acute cholecystitis not suitable for emergency cholecystectomy. Results were very similar in terms of efficacy, safety, and feasibility for both techniques, thus confirming prior findings of uncontrolled case reports and pilot studies [8,9]. As with all novel interventional procedures, the caveat of generalizability remains, as EUS -GBD was performed by a single expert operator. Improved post-procedure pain during the short time between drainage and cholecystectomy appears a small gain for EUS -GBD to compete with the widely available, time-honored, and low-cost alternative of PTGBD. However, this trial establishes short-term safety and efficacy of EUS -GBD and paves the way for EUS -GBD as a long-term palliative measure in definitive nonsurgical candidates (as opposed to those temporarily unfit for cholecystectomy). Long-term internal transmural stenting raises the question of stent dysfunction caused by occlusion or migration. The study presented next provides preliminary data to help address this question [10].
Clinical evaluation of a novel lumen-apposing metal stent for endosonography-guided pancreatic pseudocyst and gallbladder drainage (Itoi et al., Gastrointest Endosc 2012 [10]) ! EUS-guided transenteric drainage of extraintestinal collections by using tubular stents (plastic stents or SEMS) is increasingly replacing surgical or percutaneous drainage. However, tubular stents are prone to migration and leakage. A lumen-apposing SEMS has been developed recently with the aim of minimizing these risks (Axios; Xlumena Inc., Mountain View, California, USA). It is a fully-covered, saddle-shaped, 10 -15-mm diameter nitinol SEMS with bilateral anchor flanges. This design aims to hold tissue layers in apposition, allowing fistula formation between nonadherent extraintestinal fluid collections and the gut lumen. Preliminary reports in animal models and ex vivo benchtop testing are encouraging [11]. The Axios delivery catheter is attached to the echoendoscope working channel in the same way that FNA needles are attached, allowing precise step-by-step deployment. Itoi et al. tested the efficacy of Axios stents for EUS-guided drainage of pancreatic pseudocysts and gallbladders. In their retrospective study, 15 patients from a tertiary center with symptomatic pseudocysts underwent 12 transgastric and 3 transduodenal EUS-guided drainage procedures, respectively. Another five nonsurgical patients with acute cholecystitis underwent four cholecystoduodenostomies and one cholecystogastrostomy, respectively. Axios stents were deployed under combined EUSfluoroscopic guidance. Standard or ultraslim endoscopes were passed through the stents and into target collections to perform endoscopy-guided interventions, including biopsy, necrosectomy, lavage, or stone removal. Stents were scheduled for removal after a median of 35 days. All stents were successfully deployed and removed. Resolution of acute cholecystitis and pseudocysts was observed in all cases after one single drainage procedure. There were no recurrences after a median follow-up of 11.4 months for pseudocysts and 9 months for cholecystitis. Of the 20 Axios stents, 19 were found to be in place and patent at the time of removal, with one single case of asymptomatic intragastric migration. No complications ensued, except for three cases of self-limited oozing at the time of removal. In 2012, five additional uncontrolled series reported EUS-guided drainage of pancreatic fluid collections with conventional tubular SEMS in 65 patients [12 -15] and with Axios in 9 patients [16] (• " Table 1). Although technical and clinical success rates were high, complications arose in up to 33 % with tubular SEMS. Another pilot trial on EUS -GBD with Axios in 13 nonoperable acute cholecystitis patients was reported [17]. Stents were successfully inserted by a predominantly transgastric route in 11/13 (85 %), all with clinical resolution. Similar to the approach used by Itoi et al., direct cholecystoscopy was carried out in 6/11. Stents were left in situ, with no recurrences over a median of 100 days (range 24 -210 days). In conclusion, in the first-reported clinical experience with lumen-apposing stents, feasibility and high technical and clinical success were shown. However, this is a pilot, observational, and retrospective study with small patient numbers. Further larger and controlled studies are needed to validate the potential benefits of this novel stent, especially in terms of long-term patency and safety. The growing number of reports on transmural SEMS drainage of pancreatic fluid collections points to relative ease, reproducibility, and safety of direct EUS-guided SEMS insertion, with perceived improved outcomes over the standard approach using transmural plastic stents/nasopancreatic drains. The available evidence is, however, too preliminary, to judge whether this perception is justified.

Comparison of EUS-guided rendezvous and precut papillotomy techniques for biliary access (Dhir et al., Gastrointest Endosc 2012 [18])
! Successful cannulation is a critical step of therapeutic endoscopic retrograde cholangiopancreatography (ERCP). Failed cannulation renders ERCP useless and potentially dangerous, as prolonged attempts increase the risk of post-ERCP pancreatitis (PEP). Despite improved ERCP skills, even experts continue to have cannulation failure rates of 1 % -2 %. Precut sphincterotomy is the most commonly used salvage biliary access technique after initially failed cannulation. EUS-guided access to biliary and pancreatic ducts combined with fluoroscopic visualization is a hybrid procedure originally named endosonography-guided cholangiopancreatography [19] (ESCP). Three distinct variant ESCP approaches to drainage have been described and are increasingly being used as salvage of failed ERCP: transmural (e. g. choledochoduodenostomy), retrograde transpapillary (rendezvous), and antegrade transpapillary [20]. Transpapillary ESCP requires antegrade guide wire passage into the duodenum across the stricture and the papilla after EUS-guided needle access to the duct. In addition, antegrade ESCP involves dilation prior to direct, transluminal over-the-wire stent insertion through the puncture tract into the duct and across the stricture, where it is deployed under fluoroscopy (• " Fig. 1). At EUS-rendezvous, however, the transpapillary placed guide wire is used for retrograde ERCP cannulation, either by snaring the guide wire and pulling it up the endoscope working channel (classic rendezvous), or by trying cannulation alongside it (parallel rendezvous). EUS-rendezvous is thus considered by many to be the safest of all three ESCP approaches to drainage. Most ESCP experience to date has been gained in patients with malignant biliary obstruction and highly complex anatomy leading to failed ERCP even after precut by experts. Can this threshold be lowered for EUS-rendezvous? Dhir et al. retrospectively evaluated the safety and efficacy of biliary EUS-rendezvous as an alternative to precut after failed cannulation. A total of 58 consecutive patients with distal bile duct obstruction (75 % malignant) and failed selective biliary ERCP wire-guided cannulation underwent EUS-rendezvous over 12 months at a tertiary center in Mumbai. Treatment success and complications were compared with a historical cohort of 144 consecutive patients with naïve papillae who had undergone precut at the same unit during the 12-month period ending 4 months before the study. ERCP (± precut) and EUS-rendezvous were performed during a single-session by two different operators. The Erlangen papillotome was used for precut and a short wire system was used for EUS-rendezvous. Treatment success was significantly higher for EUS-rendezvous (57/58 patients) than for precut (130/144 patients) (98.3 % vs. 90.3 %). EUS-rendezvous had fewer and less-severe complications (two mild leaks) than precut (four PEP [one severe, three moderate], six bleeding [one moderate, five mild]), but the difference was not significant (3.4 % vs. 6.9 %). The authors concluded that EUS-rendezvous is superior to precut and have accordingly adopted it as the primary salvage approach to failed cannulation. Before we discuss the apparently quantum-leap conclusion that EUS-rendezvous is superior to precut, it should be noted that: a) this is the largest single-center study on EUS-rendezvous published to date; b) the short wire system (260 cm, 0.032 inch) is unique, as 450 cm long wires are standard for EUS-rendezvous; and c) the EUS-rendezvous success rate was much higher than previously reported. These facts stand out regardless of the comparison with precut. Dhir et al. attributed their success rate to the improved steerability afforded by the short wire, which effectively helped them to overcome antegrade passage, the limiting step of EUS-rendezvous for experts and novices alike [21]. Even if Dhir et al. had accrued experience with EUS-rendezvous prior to this study [22], a learning curve effect is unlikely to explain their impressive success rate. Current large series from high-volume centers report success rates of about 75 % for biliary EUS-rendezvous, definitely higher than 65 % during early experience [21], but still far from 98 %. Cross-over strategies after initially failed EUS-rendezvous can improve final ESCP success rates. Iwashita et al. tried ERCP immediately after EUS-cholangiography in 7 out of 11 failed EUS-rendezvous cases, with cannulation in 4, raising their success rate from 73 % (29 /40) to 83 % (33 /40) [23]. Similarly, Shah et al. initially succeeded in only 74 % of attempted biliary EUS-rendezvous procedures, but reached 85 % (58 /68) for biliary ESCP after cross-over [24]. Patient selection and definition of treatment success in this study warrant further comment. Favorable patient selection may have contributed to improved outcomes of EUS-rendezvous in this series. Only patients with distal obstruction and a common bile duct of > 8 mm were included. This profile is different from the one in the Shah series in which postoperative anatomy and/or proximal biliary obstruction dictated the need for intrahepatic access in up to 79 % [24]. EUS-rendezvous success rates appear to be lower with intrahepatic than with extrahepatic access (50 % vs. 80 % on average [23,25]), whereas complication rates appear to be higher [25] (17 % vs. 8 %). In another comparative study by the Mumbai group including 35 cases [26], the difference in complication rates between intra-and extrahepatic EUS-rendezvous was highly significant (70 % vs. 5 %). The stringent inclusion criteria in the current study probably explain why only 58 patients underwent EUS-rendezvous, whereas during a similar period 144 controls underwent precut. In other words, in this retrospective study, only "rendezvous-friendly" failed cannulations were selected for comparison against precut. Treatment success was defined as completion of endotherapy in a single session, leaving precut at comparative disadvantage. Overall, 10 % of patients in the precut group required a second session, with a final success rate of 95.8 %no different from the initial success rate in the EUS-rendezvous group. The unique short-wire system might also raise concerns about reproducibility. However, from a practicing endosonographer's standpoint, the minute procedural details provided by Dhir et al. warrant full attention [18,22]. Limitations aside, their study provides good evidence to support EUS-rendezvous as salvage of failed cannulation in adequately selected patients when the required expertise is available. In 2012, 401 new ESCP cases were reported in 10 different series, including the first (small) randomized controlled trial comparing EUS-choledochoduodenostomy to percutaneous drainage [27], the first multicenter study [21], and the largest single-center series to date [24] (n = 95). If anything, this tells us that this required expertise is fast becoming available. Fig. 1 Antegrade transpapillary endosonography-guided cholangiopancreatography drainage in a patient with a type II duodenal stricture and prior placement of duodenal self-expandable metal stent. a Extrahepatic bile duct access from the distal antrum for cholangiography and guide wire placement; b transpapillary guide wire passage, puncture tract dilation; c stent insertion through the puncture tract and stent deployment monitored by fluoroscopy.