First Randomized Controlled Trial of Flexible Endoscopic Mucomyotomy of Zenkers Diverticulum (ZD): APC Versus Needle-Knife Cutting

Abstracts

patients treated with RFA at our centre starting July 2005, under IRB approved protocols. Enrolled patients had BE %12cm with HGD/EC. Non-flat lesions were removed with ER prior to RFA. Exclusion criteria: cancer OT1sm1 or Nþ disease on EUS. Primary circumferential RFA was performed using a balloon-based catheter, secondary focal RFA was performed with an endoscope-based catheter (BARRX Medical). Primary RFA was performed 6 weeks after ER, followed by secondary RFA every 2 months until clearance of BE was confirmed endoscopically by inspection with narrow-band imaging, and histologically by biopsies. Thereafter, follow-up endoscopy and biopsy was performed at 2, 6, and 12 months and then annually. Results: 73 patients were included (59M, mean age 65yrs, median BE 5cm). 57/73 patients (78%) underwent ER (20 en-bloc, 37 piecemeal) prior to RFA revealing EC (nZ32), HGD (nZ20), LGD (nZ4) or non-dysplastic BE (nZ1). The worst histological grade of residual BE prior to RFA was HGD (nZ42), LGD (nZ24), or non-dysplastic (nZ7). By November 2008, 11 patients are still under treatment, while 62 patients have completed treatment (results of 44 patients have been published). Complete histological eradication of dysplasia and intestinal metaplasia (IM) was achieved in 59/62 patients (95%) after a median of 1 (IQR 1-1) circumferential RFA and 2 (IQR 1-2) focal RFA sessions, and additional ER in 5 patients. There were 3 protocol failures: 2 patients had persisting dysplasia (5%), in 1 patient (2%) treatment was ceased due to poor mucosal healing after RFA. Complications following RFA: asymptomatic non-transmural laceration at an ER-scar after circumferential RFA (nZ7), dysphagia (nZ5), melena (nZ1). No lacerations or dysphagia occurred in patients without prior ER. Seventeen months (IQR 14-33) after the last treatment, no dysplasia had recurred. In one patient a 1mm BE island was identified at 12 months follow-up. Eight patients had focal IM detected immediately distal to the neo-Z-line at a single FU. Of 2515 biopsies obtained from neosquamous epithelium during any follow-up, 2 showed buried IM (0.08%). Conclusions: RFA of BE-HGD/EC with or without prior ER of visible lesions is effective in achieving complete eradication of dysplasia and IM (95%) without serious adverse events.

W1340 Ex-Vivo Imaging of Esophageal Neoplasm Using Autofluorescence Microscopy Bevin Lin, Shiro Urayama, Ramez M. Saroufeem, Stavros G. Demos Background: Esophageal carcinoma carries a significant mortality rate with increasing U.S. incidence. Emerging photonic techniques including OCT, nonlinear microscopy, and confocal endomicroscopy show promise for disease diagnosis. Extending our previously reported multimodal esophageal imaging using autofluorescence (AF) and polarization light scattering to localize suspicious regions, we implement high-throughput, hyperspectral microscopy to provide realtime histology without contrast agents using AF imaging under UV excitation. This system provides images of epithelial microstructures that suggest esophageal neoplastic activity. Methods: IRB-approved human esophageal specimens were collected. Biopsies were imaged singly under 266 and 355 nm excitation. Spatial resolution was approximately 1 mm. After imaging, biopsies were placed in formalin for histopathological evaluation. Results: Cell junctions and nuclei were visible in squamous mucosa of Fig(a). Fig(b) highlighted abnormal vascular pattern of squamous cell carcinoma. Fig(c) displayed a characteristic normal glandular cell pattern. Darker groups of cells in Fig(d) appeared to represent goblet cells, consistent with Barrett’s esophagus. Microstructure distortion was identified in Figs(e, f). Changes of nuclear and abnormal cellular pattern of mucosa in Fig(e) corresponded to dysplastic region. Adenocarcinoma, Fig(f), demonstrated a disorganized villiform pattern of the abnormal epithelium. Conclusion: Benchtop system assessed cell morphology related to esophageal neoplastic development and progression without optical sectioning. Adaptation of this technology into a prototype endoscope probe for standard endoscopy implementation may provide real-time microscopic detail following identification of suspicious regions. Further studies investigating dysplastic lesions using this technology are ongoing.

surveillance is indicated. Esophago-gastro-duodenoscopy (EGD) is the current gold standard for evaluation of esophageal varices. Recently numerous studies have evaluated the role of esophageal capsule endoscopy (ECE) as an alternative to EGD for screening of varices. Aim: To perform a Meta analysis to evaluate the overall pooled sensitivity and specificity of ECE in detecting esophageal varices when compared to traditional EGD. Methods: Using the key words ‘Cirrhosis’, ‘variceal screening’, ‘capsule endoscopy’, ‘Gastro intestinal endoscopy’, ‘digestive endoscopy’ and ‘esophageal varices’ with various combinations using the Boolean operators AND, OR, searches were conducted in pubmed, Cochrane library, EMBASE and ISI web of science. Manual searching of referenced articles was also done. Initial search yielded 270 articles. Our inclusion criteria were 1) Studies that compared ECE and EGD for screening or surveillance of esophageal varices, 2) Use of PILLCAM ESO for capsule endoscopy. A total of 15 articles were initially shortlisted and 8 studies were selected for inclusion in the final data analysis. Results: 500 patients from 8 studies were included in the analysis. Total of 5 patients were identified as capsule failures. Analysis was done using metadisc statistical package v 1.4. There is no pooled odds ratio as we are evaluating the pooled sensitivity and specificity. Tests of heterogeneity revealed that the tests were heterogeneous shown by the high I2 values. Fixed and random effects model were used to calculate pooled sensitivity, specificity, positive and negative likelihood ratios. Pooled sensitivity and specificity for ECE for the diagnosis of esophageal varices were: SensitivityZ 0.83(0.79-0.87), SpecificityZ 0.77(0.70-0.83), Pooled negative likelihood ratioZ 0.28 (0.15-0.54), Pooled positive likelihood ratio Z 2.88 (1.06-7.88), Conclusion: The overall pooled sensitivity, specificity and likelihood ratios show ECE to be inferior to traditional EGD for the diagnosis and surveillance of varices. Large scale prospective studies are needed to further evaluate the role of ECE in this setting. Advancements in capsule technology and the learning curve associated with a new technology may make it a future choice for variceal screening and surveillance.

W1342 Endocytoscopy in Barrett’s Esophagus Observer Variation Study: Proposal for a New Classification System Yutaka Tomizawa, Ngozi I. Okoro, Ganapathy A. Prasad, Louis-Michel Wong Kee Song, Navtej Buttar, Lynn S. Borkenhagen, Kelly T. Dunagan, Lori S. Lutzke, Kenneth K. Wang Background: Endocytoscopy has been applied to detect lesions in the gastrointestinal tract and has the potential to assess histological changes. However, the interpretation of cellular and nuclear changes may be subject to the same interobserver variability as histology. We have simplified an existing system for evaluation. Aim: To assess a new classification of dysplasia in endocytoscopic images in BE and evaluate the interobserver variability. Method: Patients undergoing endoscopic mucosal resection were assessed using endocytoscopy. Lesions targeted for resection were endoscopically apparent areas of dysplasia or cancer in patients with known high grade dysplasia. Staining of the surface was done with 20% acetylcysteine and 1% methylene blue. Each specimen was imaged with a flexible catheter-type contact video endoscope (EndoCytoscope XEC 120, Olympus Tokyo Japan) with magnification x1100. These images were recorded for analysis and comparison. We modified previously proposed ECA (endocytoscopic atypia) classification into four categories (mECA). Our proposed mECA classifications included cytoplasm-rich cells with a rhomboid shape in a regular pattern (mECA1), increased cell number and different-sized nuclei/cells (mECA2), cell concentration and increased nucleus-cytoplasm ratio is high and high dense chromatin and nucleus fission are prominent (mECA3), cells of various sizes are arranged irregularly, with blurred and enlarged nuclei (mECA4). Endocytoscopy naive gastroenterology fellows were selected to participate. After receiving a didactic session about endocytoscopy and viewing an atlas about each category, we showed randomly assigned unknown-diagnosis pictures and asked the observers to classify the images. Result: A total of 22 clinical fellows participated in the interobserver valiability study. Under our new classification, overall accuracy for correct diagnosis was 81.8%. Accuracies for each category are 97.7% for mECA1, 63.6% for mECA2, 70.5% for mECA3 and 95.5% for mECA4. If we combine mECA2 and mECA3 as diagnosis of any dysplasia, the accuracy would be nearly perfect (98.9%). Conclusion: This is the first report regarding the endocytoscopic classification for BE. We propose a new classification with excellent accuracy for diagnosis of squamous epithelium and esophageal adenocarcinoma and dysplasia. For diagnosis of BE with or without dysplasia, it had a lower but still acceptable accuracy. This fact probably reflect some of the dilemma that exists with pathological interpretation of dysplasia in BE.

Fig. 1: a) squamous musoca, b) squamous cell carcinoma, c) squamocolumnar mucosa, d) Barrett esophagus, e) focal high grade dysplasia, f) adenocarcinoma.

W1341 Sensitivity of PillCam ESO for Detection of Esophageal Varices; A Meta Analysis Praveen Guturu, Sashidhar Sagi, Sathya Jaganmohan, Gagan Sood

W1343 First Randomized Controlled Trial of Flexible Endoscopic Mucomyotomy of Zenkers Diverticulum (ZD): APC Versus Needle-Knife Cutting Jean Pierre Charton, Brigitte Schumacher, Pablo E. Verde, Horst Neuhaus

Introduction: Current guidelines recommend screening for esophageal varices at diagnosis of cirrhosis and depending on the findings, treatment and subsequent

Background: Flexible endoscopic treatment has been established as a minimally invasive method for treatment of symptomatic ZD. Randomized controlled trials of

AB346 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 5 : 2009

www.giejournal.org

Abstracts

different endoscopic methods have not been reported, yet. Aim: To compare the efficacy and safety of APC versus needle knife incision for mucomyotomy. Methods: 40 consecutive patients (pts.) with symptomatic ZD referred to our unit from 2/06 to 4/08 were randomized to APC or needle-knife (NK) mucomyotomy. Primary end point was the number of sessions necessary to achieve symptom relief and pts. satisfaction. Before treatment and during follow-up all pts. were asked for evaluation of symptoms (frequency/intensity of discomfort) by numeric analogue scales (NAS). Results20 pts. were treated by APC and 20 by NK. There were no statistical differences between the 2 groups concerning gender/age or size of the diverticulum (mean size 3.1cm, range 1.2-7.2cm). 3 pts. were excluded from the study (2 pat. died during follow-up due to cardiovascular reasons; 1 pat. was lost to f/u). All of these 3 pts. had been randomized to the APC group. In 2 pts. of the APC group the method had to be changed to NK during therapy, because it was technically impossible to treat these pts. by APC. These cases remained in the original group for analysis. After a mean f/u of 14 months (6-32) 5/17 pts. of the APC group and 3/20 pts. of the NK group were treated more than once (pZ0.428; APCgroup: 12 pts. with 1 session., 5 pat. with 2 sessions; NK group: 17 pts. with 1, 2 with 2, 1 with 3). Concerning clinical symptoms no statistical significant differences were documented. Both groups of pts. reported of comparable improvements in terms of frequency and intensity of discomfort (qualities: drinking, eating smooth/ solid food). We diagnosed 4 complications in the APC-group (23%; 1 emphysema, 3 esophageal stenosis) and 1 complication in the NK group (5%; 1 emphysema). Due to the relatively small sample size the difference was not significant (pZ0.159) even if only esophageal stenoses were considered (pZ0.23). All complications could be conservatively or endoscopically treated. 3 pat. (each with a minimum of 2 sessions) are still under investigation. Conclusion: Flexible endoscopic mucomyotomy of symptomatic ZD by APC or NK is safe and effective. This randomized controlled trial revealed no significant statistical differences in terms of number of sessions necessary to achieve symptom relief or improvement and complications, respectively. There was a tendency for less sessions necessary for resolution of dysphagia and less complications in the needle-knife-group.

W1344 Endoscopic Resection for Mucosal Esophageal Cancer - Is There a Difference in Outcome Between Mucosal Barrett’s Cancer and Mucosal Squamous Cell Cancer? Oliver Pech, Hendrik Manner, Andrea May, Juergen Pohl, Christian Ell Background: Endoscopic resection (ER) with curative intent is considered to be a safe and effective alternative treatment to radical surgery in case the neoplasia is intraepithelial or limited to the mucosal layer in both, Barrett’s cancer (BC) and squamous cell cencer (SCC). However, up to now there is now study comparing both entities regarding the outcome after ER. Objective: To compare the long-term safety and efficacy of ER for mucosal SCC and BC in a large Western population. Design: Retrospective uni-center analysis. Setting: Tertiary-referral center between October 1996 and September 2007. Patients: Subjects with mucosal SCC and BC. Exclusion criteria were submucosal or more advanced cancers and follow-up !12 months. Patients with BC were selected from our database and matched to the patients with SCC regarding age, gender, tumor stage (mucosal cancer) and treatment method (ER). Interventions: ER in 183 patients (85 patients with SCC and 98 patients with BC). Main Outcome Measurements: Complete remission of neoplasia and long-term survival. Results: The mean follow-up in those patients with SCC and BC was 50.1 and 49.1 months, respectively. Complete remission was achieved in BC (95/98) significantly more often than in SCC (76/85) (97% vs 89%; pZ0.04). SCC had a significant larger diameter than BC (p!0.0001) and more ER per patient were performed in SCC than in BC (2.67 vs 2.41; pZ0.007). Recurrence occurred in 24% (20/85) of patients with SCC and in 13% (13/98) with BC (pZ0.72). Significantly more patients with SCC died during follow-up than those with BC (19% vs 7%; p!0.0001). There were 4 tumor-related deaths in patients with SCC and none in BC (pZ0.045). Complications occurred more often in the SCC-group (34% vs 13%; pZ0.0005; esophageal strictures 31% vs 2%; pZ0!0001). However, we observed more bleedings during ER of BC than of SCC (4% vs 11%; pZ0.04). Limitations: Uni-centric retrospective trial. Conclusions: Initial and long-term results of ER are significantly better in patients with BC than with SCC. ER in SCC is associated with a significantly higher stricture rate, however bleedings are more frequently observed after ER of BC.

W1345 Long-Term Results of Endoscopic Radial Incision and Cutting (ERIC) Method for Intractable Esophageal Stricture Manabu Muto, Tomonori Yano, Keiko Minashi, Shinya Tsuruta, Shuko Morita, Horimatsu Takahiro, Yasumasa Ezoe, Tsutomu Chiba Background: Endoscopic balloon dilation (EBD) has been carried out for the stricture after the operation or EMR/ESD or chemoradiotherapy (CRT) for esophageal cancer. However, some patients (pts) suffered from long-term and severe dysphasia due to intractable stricture that was refractory to the repeated EBD, and there has been no effective treatment. To relieve the symptom of severe dysphasia, we developed a new treatment, endoscopic radial incision and cutting

www.giejournal.org

(ERIC) method. Aim: We examine the safety, short-term results, and long-term result after 6 months of the ERIC treatment in the pts with severe esophageal stricture. Pts and method Pts who did not improve the dysphasia even after repeated EBD were indicated for ERIC. ERIC was carried out as follows: (1) The stricture part was radically incised by using insulation tip (IT) knife endoscopically. (2) The line which connects the esophageal lumen on the oral side and lumen on the anal side is assumed, and is incised along the line. (3) The incision part is sliced off by IT knife. (4) After incision, EBD is carried out at the frequency of 1-2 times a week to prevent the re-regurgitation until the cutting surface become scar. Results We performed ERIC for 29 pts between May, 2006 and August, 2008. Median Age: 67-year-old (range 33-84) and male to female ratio was 24: 5. Cause of stricture: post EMR: 7, post operation: 19, post salvage treatment after CRT: 3. Median follow up period was 7 months (range 4-175) and median EBD times before ERIC were 13 (range 4-62). Before ERIC, 15 pts could eat solid food, however, they suffered from difficulty of swallowing and took long time to finish the meal. The remaining 14 pts could eat only semi-solid or liquid food. After ERIC, 97% (28/29) of the pts could eat solid food without symptom of dysphasia. Regarding the size of the esophageal lumen, 86% (25/29) of the stricture improved to be passed through by the conventional endoscope. Median procedure time was 15 minutes (range 5-40). No serious adverse events during ERIC were observed. Median hospital stay was 4 days (3-11). At the time of 6 month after ERIC, 22 pts can be evaluated and 20 pts (90%) could eat solid food. With a median times of preventive EBD of 11 (range 1-71), 55% (15/29) pts completely relived the symptom of dysphasia and become unnecessary for EBD. While 7 pts (24%) required ERIC again after initial ERIC, their symptom was relieved during the period until re-stenosis. Conclusions ERIC method for the intractable stricture is the effective and safety method. By demonstrating the validity of this method, ERIC could be a completely new medical treatment for patients with severe stricture.

W1346 Endoscopic Screening of Esophageal Cancer in Patients with Head and Neck Cancers: A Prospective Study By Comparison of Narrow-Band-Imaging with High-Magnification and Conventional Endoscopy Ching-Tai Lee, Chi-Yang Chang, Chi-Ming Tai, Jau-Chung Hwang, Wen-Lun Wang, Jawtown Lin Background and Aims: Esophageal cancers coexistent with patients with head and neck cancers are not unusual. Early detection of esophageal cancer in these patients may alter treatment planning and thus further improve the survival. Recent advances in narrow-band-imaging (NBI) in endoscopy permits adjustment of reflected light, enhancement of the vascular pattern, and improvement of the detection rate of early-stage mucosal malignancies. In this prospective study we aimed to evaluate the prevalence of concomitant esophageal cancers in patients with head and neck cancers. We also compared the detection rate of NBI with highmagnification and conventional endoscopy in screening these esophageal cancers. Patients and Methods: A prospective study was undertaken in single endoscopy center in South Taiwan from April 2008 to November 2008. A total of 50 patients with documented head and neck cancer were enrolled. Demographic data and risk factors were recorded. All patients underwent a meticulous endoscopic examination of the esophagus with conventional white light system, followed by switching to NBI system. If a brownish area under NBI was detected, highmagnification was applied to determine the microvascular pattern in the lesion. Biopsy for all suspicious neoplastic lesions by conventional endoscopy and/with/or NBI with high-magnification was done for final pathology. Results: A total of 19 endoscopic suspected lesions were obtained from 15 patients using conventional endoscopy, while there were 30 endoscopic suspected lesions obtained from 20 patients using NBI with magnification. All suspicious lesions found by conventional endoscopy could be also detected by NBI with magnification. NBI system could detect more superficial lesions than conventional endoscopy (20 vs. 9 lesions). After histological analysis, NBI with magnification found more neoplastic lesions than conventional endoscopy (24 lesions in 14 patients vs. 15 lesions in 12 patients), especially in patients with high grade dysplasia (5 lesions in 4 patients vs. 0 lesion). Totally 14 patients were diagnosed as squamous cell carcinoma of esophagus in 50 patients with head and neck cancers, resulting in a prevalence of 28%. The overall concordance in interpretation between these two methods was 81.7%. Conclusion: The prevalence of esophageal neoplasia reached 28% in patients with head and neck cancers, which was higher than previous study. NBI with magnification significantly increased the detection rate of suspicious neoplastic lesion, especially the superficial neoplasia. Periodical endoscopic surveillance of esophagus in patients with head and neck cancers is warranted.

W1347 Prospective Single-Center Cohort Study for the Evaluation of the Long-Term Clinical Efficacy of Flexible Endoscopic Diverticuloscope-Assisted Zenker’s Diverticulotomy Federico Iacopini, Vincenzo Perri, Alessandra Bizzotto, Andrea Tringali, Michele Marchese, Massimiliano Mutignani, Guido Costamagna

Volume 69, No. 5 : 2009 GASTROINTESTINAL ENDOSCOPY AB347