Abstracts

36, abscess 15, necrosis 9). Technical success was 95% (57/60): failures were due toinability to drain thick necrotic PFC in 2 and inability to visualize PFC in 1.Treatment success was 93% (53/57): 3 abscesses and 1 necrosis required surgery. Nocomplications were encountered. There were no significant differences in patient/clinical characteristics between group A and B patients. While there was nodifference in technical or clinical outcomes, median procedural duration wassignificantly shorter for group B than group A patients (25 vs.70 min; p ! 0.001).After 25 procedures, procedural duration for performing EUS-D of PFC was morelikely to be ! 30 minutes (adj. OR Z 11.6; p Z 0.01) even after adjusting forpatient age, serum albumin, type/location/size of PFC, drainage modality (stent vs.stent plus drainage catheter) and site of endoscopic access in GI tract forestablishing drainage. Conclusions: In this study, the largest reported to-date, EUS-guided drainage of PFC was performed safely using the graded dilation techniquewith a successful outcome in a majority of patients. Learning curve, with regard toprocedural duration, improved significantly after the first 25 cases.

M1494

EUS-Guided Drainage: An Alternative to Surgical and

Percutaneous Approaches for Management of Multiple,

Non-Communicating Peripancreatic Fluid CollectionsShyam Varadarajulu, Jeanetta BlakelyBackground: Management of multiple peripancreatic fluid collections (PFC) isa therapeutic challenge, particularly in patients with disrupted main pancreatic ductas pancreatic stenting is not feasible. Also, endoscopic trans-mural drainage maynot be feasible as none or only few of the PFC may cause definitive luminalcompression (LC). Under such circumstances, management is relegated topercutaneous/surgical drainage. Although the role of EUS for drainage of PFCwithout LC is well known, there are no published reports on use of this modalityfor drainage of multiple PFC. Aim: Evaluate the role of EUS in drainage of multiple,non-communicating PFC. Methods: Prospective study of all patients whounderwent EUS-guided drainage of PFC over a 42-month period. Prior to EUS, allpatients underwent contrast enhanced abdomen CT and ERCP was attempted forpancreatic stenting. In symptomatic patients with multiple PFC that were non-communicating, not causing LC, and large (O 6 cm), EUS-guided drainage of PFC inall locations was undertaken. At EUS, the PFC were accessed using a 19-gaugeneedle and after passage of a 0. 035 inch guidewire, sequential dilation of the tractwas performed and 7 Fr/10 Fr stents or drainage catheters were deployed. Clinicalfollow-up and CTwas undertaken at 6-weeks to assess treatment response. Results:Multiple, large, non-communicating PFC was encountered in 6 (3 pseudocyst, 2abscess, 1 necrosis) of 60 patients (10%; 95% CI [4.3%, 20.4%]) who underwentEUS-guided drainage. Etiology of pancreatitis in the 6 patients (3 female, mean age60.5yrs [range, 41-70]) were idiopathic in 4, gallstone 1 and ETOH 1. ERCP revealedcomplete duct disruption in all 6 patients. Three patients with pseudocyst had PFCat 3 locations and others at 2 locations. The procedure was technically successful inall patients (100%). A total of 15 PFC sites were drained under EUS-guidance inthese 6 patients. Sites of access were GE junction in 1, gastric cardia/fundus (6),greater curvature (2), antrum (2), duodenum (3), and Roux-en-Y loop in (1). Themean procedural duration was 60 minutes (range, 47-76). No complications wereencountered. At follow-up, 4 (3 pseudocyst, 1 abscess) of 6 patients (66%) hadcomplete symptom relief and resolution of PFC; others experienced persistentsymptoms with residual fluid collection requiring surgical drainage. Conclusions:EUS-guided drainage of multiple, non-communicating peripancreatic fluidcollections is technically feasible, safe, obviates the need for surgery or placementof multiple percutaneous drainage catheters, and leads to a successful clinicaloutcome in two-thirds of patients.

M1495

EUS Versus Surgical Cyst-Gastrostomy for Management

of Pancreatic PseudocystsShyam Varadarajulu, Tercio Lopes, Charles M. Wilcox,Ernesto R. Drelichman, Meredith L. Kilgore, John D. ChristeinBackground: Although EUS-guided cyst-gastrostomy is increasingly performed,there are no studies comparing the clinical outcomes and cost-effectiveness of thisapproach with ‘gold standard’ surgical cyst-gastrostomy. Aim: Compare the clinicaloutcomes of EUS-guided cyst-gastrostomy with surgical cyst-gastrostomy formanagement of patients with pancreatic pseudocysts and perform a cost analysis ofeach treatment modality. Methods: This is a retrospective case-control study ofconsecutive patients with uncomplicated pancreatic pseudocysts who underwenteither surgical or endoscopic cyst-gastrostomy over a 3-yr period at a tertiaryreferral center. Individual patients who underwent surgical cyst-gastrostomy werematched with two patients who underwent EUS-guided cyst-gastrostomy duringthe same time period for age, etiology of pancreatitis, and size of the pseudocyst.All procedures were performed by one experienced pancreaticobiliary surgeon andone experienced endosonographer. Patients with pancreatic abscess or necrosiswere excluded. Main outcome measures were rates of treatment success,complications and reinterventions, length of hospital stay (post-procedure), andcost associated with each treatment modality. Treatment success was defined asresolution of clinical symptoms and pseudocyst on follow-up imaging at 6-weeks.Results: Ten patients (6 male, median age 41yrs, mean pseudocyst size 6.2 cm

AB226 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 5 : 2008

[range, 5-16]) who underwent surgical cyst-gastrostomy were matched with 20patients who underwent EUS-guided cyst-gastrostomy. There were no significantdifferences in demographic data, comorbidity, and clinical characteristics betweenboth groups. While there were no significant differences in rates of treatmentsuccess (100% vs. 95%; p Z 0.36), procedural complications (none in both groups)or reinterventions (10% vs. 0%; p Z 0.13) between surgery vs. EUS-guided cyst-gastrostomy, the mean length of hospital stay for EUS-guided cyst-gastrostomy wassignificantly shorter than for surgical cyst-gastrostomy (2.6 vs. 6.5 days; p Z 0.008).The average total cost per case for the EUS-guided cyst-gastrostomy approach wassignificantly less when compared to surgical cyst-gastrostomy ($7943 vs. $14,815,p Z 0.01) with the total cost savings being $6872 per patient. Conclusions: Whenavailable, EUS-guided cyst-gastrostomy should be the first-line treatment approachfor patients with uncomplicated pancreatic pseudocysts as the procedure isless costly and is associated with shorter length of hospital stay than surgicalcyst-gastrostomy. There was no significant difference in clinical outcomes betweenboth treatment modalities.

M1496

One-Step, EUS-Guided Drainage of Pancreatic Pseudocysts:

Experience in 52 PatientsColin Mckay, Simon Denley, Ross CarterIntroduction: In the treatment of pancreatic pseudocysts endoscopic treatment isnow standard. Increasingly, trans-gastric drainage is being used in patients withacute pancreatitis and patients with more complex cysts containing debris. Wereport our experience in 52 patients using a one-step, two stent procedure underEUS guidance. Methods: All patients had persistent, symptomatic cysts. Aetiologywas acute pancreatitis in 34 and chronic pancreatitis in 18 patients. 17 cysts wereinfected. ERCP or duct stent placement was either unsuccessful (8 cases) ordemonstrated no communication (31 cases). In 10 cases ERCP was not attempted.In the remaining 3 cases, cysts persisted despite pancreatic duct stenting. Cystdrainage was carried out under EUS control using a therapeutic echoendoscope(Pentax EG-3830UT) with a 3.8 mm working channel and bridge. Two or morepigtail stents were inserted after entering the cyst with a Wilson-CookCystotomeTM and performing balloon dilatation (oesophageal balloon) of thecystgastrostomy to 12-15 mm. Results: Cysts resolved at the first procedure in 37cases, with one patient undergoing 24h of lavage via a naso-cyst catheter. A further5 patients underwent a second procedure, 3 for placement of a naso-cyst catheterand lavage, 2 for further dilatation of the cyst-gastrostomy. All 5 resolved aftera second procedure. One patient underwent 4 procedures, a naso-cyst catheter atthe first followed by 3 further dilatations. In 3 patients formal open surgery wasrequired, one 2 months later for incomplete drainage, the other two for procedurerelated complications. Two patients died following surgery for persistent, infectedcysts. A third death occurred due to liver failure following successful cyst drainageand a fourth died approximately 1 year after successful cyst drainage froma pancreatic tumour. Conclusions: One-step EUS-guided drainage with thedeployment of two or more stents is an effective treatment with 43 of 52 patientshaving complete cyst resolution by endoscopic means only. Complications alloccurred early in this study indicating a learning curve, but these can be minimisedby appreciating the need for further intervention and the use of naso-cyst lavage inmore complex cysts.

M1497

Effectiveness and Safety of EUS Guided Pancreatic Pseudocyst/

Abscess Drainage: A Systematic Review of Prospective StudiesPriyanka Tiwari, Nadim S. Jafri, Praveen K. Roy, Mainor R. AntillonBackground: EUS guided pseudocyst and abscess drainage has been reported sinceit was first described in 1992. Overall success rates and complications of thisprocedure are however not known and it is unclear if this treatment is definitive.We conducted a systematic review to evaluate the effectiveness and safety of EUSguided drainage of pancreatic pseudocysts and abscesses. Method: We performeda systematic search of Medline, EMBASE, and Cochrane Central Register of ControlledTrials without language restriction for prospective studies evaluating the role of EUSguided drainage of pancreatic pseudocysts/abscess. We also hand searched greyliterature from major gastrointestinal meetings. Standard forms were used to extractdata regarding study design, cyst/abscess size, success rates, recurrence, surgicalintervention, mortality and complications by two independent reviewers. Result: Sixstudies (156 patients) met our predefined eligibility criteria. The sample size of thestudies ranged from 8 to 46 patients. Patient age ranged from 9-81 years. Time ofintervention from time of diagnosis was less than 4 weeks. The length of follow-upperiod after the therapeutic intervention ranged from 3-468 weeks among the studies.The etiology of cyst/abscess formation reported in the studies were: alcohol, acutepancreatitis, gallstones and other causes. The size of the pseudocysts/abscess rangedfrom 4-14 cm in all the studies. The pancreatic body was the most common location(22-76% of patients), followed by the head of the pancreas (12-52%) and pancreatictail (8-34%). In two prospective studies, 3-8% patients were also found to have gastricvarices at the time of endoscopy. An endoscopic bulge was seen in 3-12% patients (3studies). Overall, the success rate for EUS guided drainage of pancreatic pseudocystsand abscesses varied from 82%-100% among the studies. The most commoncomplications from the endoscopic intervention were bleeding, perforation, stentmigration and stent occlusion. The rate of recurrence of the pesudocyst/abscess

www.giejournal.org

Abstracts

varied from 0-11% of patients in the studies. Subsequent surgical intervention wasneeded in 0-10% of the patients. There were no deaths related to the procedure atthe end of the follow up period in any of the studies. Conclusion: EUS guidedpancreatic drainage is an effective and safe definitive procedure.

M1498

Therapeutic Endoscopic Ultrasound (EUS) Access and/Or

Stenting of Pancreaticobiliary System After Failed Conventional

Endoscopic Retrograde Cholangiopancreatography (ERCP)H. Nguyen Thomas Thong, Christopher E. Lee, Charles S. Whang,Patrick B. Truong, Raman Muthusamy, Kenneth Chang, John G. LeeBackground: Pancreaticobiliary access during ERCP can fail due to duodenalobstruction, surgically altered anatomy or unsuccessful cannulation. We used EUSto obtain pancreatic or biliary access and/or stent after conventional ERCP failed.Method: Retrospective review of 18 consecutive patients undergoing EUS guidedaccess during 12 months. We assessed outcomes from medical records andreferring physicians. Results: We imaged the desired duct with linear EUS andinjection of radiographic contrast. A 0.035 inch guidewire was advanced throughthe 19g needle and manipulated across the stricture, ampulla or intrahepatics. TheEUS scope was removed and a conventional endoscope was advanced to the wire,which was grasped and pulled out. The duct was accessed retrograde over the wireif possible; if not the stent was placed through an endoscopically createdcholedochoduodenostomy (CD). ERCP failed in 18 (8 men, mean age Z 68y, range30-94y) due to malignant duodenal obstruction resistant to dilation (6), priorduodenal metal stent (2), malignant choledochojejunal stenosis post Whipple (4)or roux en Y (1), benign choledochojejunal stenosis post roux en Y (1), benignpancreaticojejunal stenosis post Whipple (1), ampulla inside diverticulum (2) andfailed deep cannulation (1). We stented retrograde over a guidewire if possible (9);otherwise through an endoscopically created CD with the distal end of the stent inthe bulb (3) or the ampulla (1). The patient with stones had duct clearance andjaundice resolved in all with technical success. The patient with pancreaticojejunalstenosis and recurrent pancreatitis resolved the stenosis and remainsasymptomatic. Two with CD underwent planned stent exchange uneventfully andone had successful planned Whipple. Initial attempt at wire passage failed in 5/18;reattempt was successful in 1 and switching back to ERCP was successful in another(83% technical success). Percutaneous biliostomy was successful in 2. We used EUSto puncture and dilate a tract to the proximal end of the afferent limb in a patientwith recurrent cholangiocarcinoma after choledochojejun-ostomy, but still couldnot locate the anastamosis even after direct intubation of the limb usinga colonoscope; gastrojejunostomy was clipped closed without complications.Further attempts at radiographic and surgical biliary access also failed. Onecomplication of pain occurred in a patient who had antegrade metal stentplacement via CD; urgent laparoscopy was unremarkable and the patient recovereduneventfully. Conclusion: Therapeutic EUS guided access and stenting of thepancreaticobiliary system is a reasonable alternative when conventional ERCP fails.

M1499

No Cytopathologist? No Problem Maintaining Yield While

Minimizing EUS Guided FNA Passes in a Community Hospital

SettingMukul Arya, Siddharth Mathur, Niket Sonpal, Philip Mathew,William L. Thelmo, Yashpal AryaTraditionally, EUS guided FNA is performed in the presence of a cytopathologist inan effort to maximize yield and reduce the number of passes. The majority of theseprocedures are performed in tertiary care centers with access to a cytopathologist.Previous studies have reported that a minimum of 7 passes for solid pancreaticlesions and 5 passes for lymph nodes provides a sensitivity of 83% and 77%,respectively. As this modality spreads to community hospitals where services ofa cytopathologist are not readily available, there may be significant reduction inspecimen yield. The purpose of our study was to maintain adequate sensitivity inthe absence of a cytopathologist in a community hospital setting. In our trial, allsamples obtained were placed directly into a fixative solution eliminating thepreparation of slides. A cell block was created and examined by a pathologist. Wehypothesized that initial passes provided the greatest yield and specimen may belost in the preparation of slides. This would potentially reduce the number ofrequired passes while still maintaining and possibly enhancing our overall yield.Limiting passes provides obvious clinical and procedural benefits with an overallreduction in cost per. A prospective study in which all patients underwent linearechoendoscopy and FNA using a 22 g , 25 g or 19 g quick core needle. A maximumnumber of 3 passes was made regardless of the site of the lesion. The specimen wasimmediately placed into 20 cc of a methanol based fixative solution and sent for cellblock preparation. The specimens were spun and the resultant cell block wassectioned and placed on slides for histopathologic interpretation. A total of 27lesions in 27 patients were sampled. Solid lesions numbered 22 and 5 were cystic. Amean number of 2.2 passes was made. The lesions by locality were 11 pancreatic, 4lymph nodes, 3 gastric, 2 biliary, 1 esophageal and 6 cystic lesions. Overall yield onsolid lesions was 90.9% (20/22) and 100% on cysts. For pancreatic lesions thesensitivity was found to be 90.8%. These preliminary results suggest yield andsensitivity can be maintained especially in solid pancreatic tumors despite the

www.giejournal.org Vo

absence of a cytopathologist. By eliminating slides we were able to limit the numberof passes (mean -2.2) which translated into a reduction in procedure time,anesthesia requirement and overall cost without compromising sensitivity. In fact,our study is the only one to date that imposes a limit of 3 passes irrespective of thetarget lesion. Larger series have reported a range of passes with maximal numbersof 18. Data is still being collected with hopes of a larger series in the near future.

T1479

Diagnosis of Intraductal Papillary-Mucinous Neoplasm of the

Pancreas By Using Peroral Pancreatoscopy (POPS)- the

Usefulness of POPS with Narrow -Band Imaging (NBI)Taro Hara, Taketo Yamaguchi, Kentaro Sudo, Kazuyoshi Nakamura,Tadamichi Denda, Takeshi IshiharaBackgroud &Aim : Peroral pancreatoscopy (POPS) contribute useful information forthe diagnosis of intraductal papillary-mucinous neoplasm (IPMN) , and we havereported the efficary of POPS with special reference to the type of protruding lesionas well as vascular image (Gastroenterology 2002). Recently, Narrow -Band Imaging(NBI) has been used in clinical procedure as new diagnostic techniques for GI-tractdiseases. POPS with NBI has made it possible to increase the visibility of capillariesand other delicate tissue surface structures The aims of this study are to determinethe usefulness of POPS (with NBI) in IPMN for differentiation of malignant frombenign disease, and to evaluate the significance of these techniques as preoperativeexamination. Subjects and Methods: One hundred and twenty-onehistopathologically confirmed patients with IPMN underwent POPS preoperatively.;POPS with NBI in 9 patients. Findings of POPS were compared with histopathologyof resected specimens. The postoperative follow-up data were analyzed. Results:Protruding lesions were detected by POPS in 73 patients. They were classified into 5groups. ;granular type (type1), fish-egg-like type without vascular images (type2),fish-egg- like type with vascular images (type3), villous type (type4), and vegetativetype (type 5). Fish-egg-like type with vascular images, villous type and vegetativetype were considered to be malignant. The sensitivity, specificity, and accuracy ofPOPS in differentiating these 3 types as malignant were 65%, 78%, and 73%,respectively. Protruding lesions were detected in 8 out of 9 patients with NBI , andemphasized image features such as mucosal structures and capillary vessels in allthese patients. The 3-year cumulative survival rate and disease-free survival ratewere extremely high at 95%, 92%. Conclusions: POPS (with NBI) results ina considerably improved differential diagnosis between malignant and benign IPMNand useful for determining an effective therapeutic approach. Especially, NBIfacilitated easy recognition of vascular image in the protruding lesion, enhancingdiagnostic value of POPS. This technique can contribute to improvements inpostoperative results.

T1480

Intraluminal Implantation of Radioactive Stents for Treatment

of Unresectable Pancreatic Cancer, Extrahepatic Bile Duct

Carcinoma and Ampullary Carcinoma: A Pilot TrialYan Liu, Zheng LuPurpose: The aim of this study was to evaluate the potential curative efficacy andsafety of intraluminal brachytherapy through radioactive stents in the palliativetreatment of bile duct, pancreatic head and ampullary carcinomas. Materials andMethods: Patients with inoperable extrahepatic bile duct (n Z 2), pancreatic head(n Z 6) or ampullary (n Z 3) carcinomas were treated by intraluminal implantationof radioactive stents designed according to a computerized treatment planningsystem. Both radioactive stents and commonly used self-expand metal (twopatients) or plastic stents (nine patients) were placed in the common bile duct(CBD) of all patients except one pancreatic cancer patient in whom the stents withradioactive seeds were placed in both CBD and the pancreatic duct. The primarymeasures for the efficacy included abdominal pain and Karnofsky performancestatus. Survival, tumor status, and global quality of life (QOL) and symptoms werealso assessed during the period of follow-up. Results: All 11 patients weresuccessively placed with a total of 16 radioactive stents. There were no life-threatening complications. Three (27.3%) patients were classified as clinical benefit

lume 67, No. 5 : 2008 GASTROINTESTINAL ENDOSCOPY AB227