- Email: [email protected]
162 The Yield of Early Follow-up Endoscopy in Patients With Initial Nondysplastic or Low-Grade Dysplasia Barrett's Esophagus
Abstracts
identified from a group of subjects who previously completed a validated symptom questionnaire. Patients were randomized (stratified by age, sex and reflux symptoms) and invited to undergo either uTNE in a mobile research van (muTNE), uTNE in outpatient endoscopy suite (huTNE) or sEGD. uTNE was performed using a portable esophagoscope with a disposable sheath. Procedure performance characteristics (completion rate, acquisition of biopsies, duration) were systematically recorded. Validated tolerability scales (0 Z none and 10 Z severe) and adverse events questionnaires were administered 1 and 30 days after procedures. Results: 459 subjects were contacted and 209 (46%) agreed to undergo study procedures (muTNE nZ76, huTNE nZ72, sEGD nZ61). Baseline characteristics were comparable among the three groups (table 1). Participation rates were numerically higher in the unsedated arms (muTNE 47.5%, huTNE 45.7%) than in the sEGD arm (40.7%) (pZ0.27). Patients with acid reflux symptoms R1/week were more likely to participate (odds ratio 2.94, 95% confidence interval 1.47, 5.88). Complete evaluation of the esophagus was comparable using muTNE (99%), huTNE (96%) and sEGD (100%) techniques. However, successful biopsy acquisition was lower in the muTNE (79%) and huTNE (83%) groups compared to sEGD (100%) (pZ0.001). Mean duration (minutes) of esophageal examination was shorter in the sEGD arm (4.7) than in the muTNE (8.0) and huTNE (8.5) groups (p!0.001). However, recovery time was much longer for sEGD (67.3) compared to muTNE (15.5) and huTNE (18.5) techniques (p!0.001). While overall tolerability for sEGD was better than muTNE and huTNE (mean score 0.4 vs. 1.9 and 2.2 respectively, p!0.001), the majority of patients who underwent muTNE and huTNE were willing to undergo the same procedure again in future (79% and 83%, respectively). No serious adverse events were reported. 16 subjects (7.6%) were diagnosed with BE (table 2). Conclusion: In this first large randomized trial evaluating novel approaches for community screening for BE, unsedated mobile van and clinic screening with TNE was feasible and effective. The patients’ visit was significantly shorter with adequate tolerability, acceptability and safety profiles. Mobile and outpatient techniques may provide a cost-effective alternative to sEGD for BE screening. Table 1. Baseline characteristics of subjects participating in the three
arms of the randomized trial (total n[209). Data presented as number (percentage) or mean (+/- standard deviation). PPI [ proton pump inhibitors. Variable Age (year) Male gender White ethnicity BMI Waist to hip ratio Education Less than high school High school or some college College/professional training Employment Employed Unemployed/homemaker Retired Marital status Married Not married Frequency of reflux symptoms R1/week !1/week None Current use PPI drugs Current aspirin use Symptom somatisation score Charlston comorbidity index
muTNE (n[76)
huTNE (n[72)
sEGD (n[61)
65 (9) 32 (42.1%) 73 (96.0%) 29.0 (5.6) 0.91 (0.09)
64 (9) 35 (48.6%) 72 (100%) 30.5 (13.9) 0.92 (0.11)
66 (9) 29 (47.5%) 61 (100%) 28.8 (5.8) 0.91 (0.09)
2 (2.7%) 37 (49.3%) 36 (48.0%)
1 (1.4%) 40 (55.6%) 31 (43.1%)
1 (1.7%) 33 (55.0%) 26 (43.3%)
23 (30.7%) 6 (8.0%) 46 (61.3%)
29 (40.3%) 6 (8.3%) 37 (51.4%)
17 (28.3%) 5 (8.3%) 38 (63.3%)
57 (76.0%) 18 (24.0%)
62 (86.1%) 10 (13.9%)
42 (70.0%) 18 (30.0%)
11 (14.5%) 26 (34.2%) 38 (50%) 15 (19.7%) 38 (50.0%) 0.8 (1.3) 0.9 (1.3)
10 (13.8%) 32 (44.4%) 27 (37.5%) 9 (12.5%) 37 (51.4%) 0.9 (1.4) 0.9 (1.4)
8 (13.1%) 28 (45.9%) 25 (41.0%) 11 (18.0%) 36 (59.0%) 0.8 (0.5) 1.3 (2.0)
Table 2. Findings suggestive of reflux-related mucosal injury. Data
presented as number of patients (percentage) Findings Normal Hiatus hernia Suspected BE on endoscopy Confirmed BE on histology Erosive esophagitis
muTNE (n[76) 58 48 12 7 12
(76.3%) (63.2%) (15.8%) (9.2%) (15.8%)
huTNE (n[72) 36 41 11 3 25
(50.0%) (56.9%) (15.3%) (4.2%) (34.7%)
sEGD (n[61) 23 34 15 6 24
(37.7%) (55.7%) (24.6%) (9.8%) (39.3%)
P value !0.001 0.67 0.37 0.29 0.003
AB114 GASTROINTESTINAL ENDOSCOPY Volume 79, No. 5S : 2014
161 Are the Gastric Folds Still the Optimal Landmark for Defining the Distal Border of a Barrett’s Esophagus in a Western Population? Dirk SchöLvinck*1,2, Osamu Goto3, Cees a. Seldenrijk4, Jacques J. Bergman2, Naohisa Yahagi3, Bas L. Weusten1,2 1 Gastroenterology and Hepatology, St. Antonius Hospital, Nieuwegein, Netherlands; 2Gastroenterology & Hepatology, Academic Medical Center, Amsterdam, Netherlands; 3Cancer Center, Keio University, School of Medicine, Tokyo, Japan; 4Pathology, St. Antonius Hospital, Nieuwegein, Netherlands Background: The definition of the esophagogastric junction (EGJ) for a diagnosis of Barrett’s esophagus (BE) is not uniform: according to Western guidelines, the EGJ is located at the top of the gastric folds (GF), assessed during expiration. In contrast, in Japan the EGJ is defined by the distal end of the palisade vessels (PV). However, PV are faint and often not clearly visible in Western patients. Narrow Band Imaging (NBI) enhances contrast in mucosal structures and vessels. The new Olympus EXERA III endoscopy system combines bright NBI with dual focal high-resolution endoscopy, and may more easily visualize PV, even in Western BE patients. Aim: To evaluate the detection rate of PV using the new EXERA III system, to quantify the discordance between the Western and Japanese criteria for the distal border of BE, and to evaluate the presence of intestinal metaplasia (IM) in this "zone of discordance" (ZoD), in Western BE patients. Methods: Consecutive Dutch patients with BE segment of R 2 cm (no dysplasia or low-grade dysplasia) on previous endoscopy were enrolled. Endoscopies were performed by a Western expert endoscopist, assisted by a Japanese endoscopist. The distance of the GF and distal ends of PV (when visible) to the incisors were determined in in- and desufflated conditions. Any difference in distance between the top of the GF (assessed during expiration) and the distal end of the PV was considered a ZoD. Two biopsies were taken from the cardia (either 1 cm below the distal end of the PV, or 1 cm below the GF in case of absence of PV), two biopsies from the ZoD when present, and 4-quadrant biopsies starting 1 cm proximal of the GF to determine the presence of IM. Results: 25 patients (19 males, median 66 yrs, median BE C5M7) were included. PV were seen in 96%. They were more often visible during inspiration (96%) than in expiration (76%; pZ0.06). PV were located in a constant position during in- and expiration. In patients with visible PV, in 63% (nZ15) a ZoD was present during expiration with the GF median 1 cm (range 1-2) proximal to the PV. In patients with a ZoD, IM was present in 6/30 (20%) of the ZoD biopsies. This was more compared to the cardia biopsies (0/30; 0%; pZ0.01) and less compared to the biopsies obtained proximal of the GF (35/60; 58%; p! 0.001). In the 10 patients without ZoD biopsies of the cardia and proximal of the GF contained IM in 0% and 63% respectively. Conclusion: In Western patients PV are frequently seen with the latest version NBI-system. PV are best observed under optimal insufflation. In a substantial subset of patients the GF and PV differ in location and IM can be detected in biopsies from this "zone of discordance", whereas no IM is detected in biopsies distal to the PV. These results question the value of the GF as the distal margin of the Barrett’s segment.
162 The Yield of Early Follow-up Endoscopy in Patients With Initial Nondysplastic or Low-Grade Dysplasia Barrett’s Esophagus Kavel Visrodia*1, Prasad G. Iyer2, Cathy D. Schleck3, Alan R. Zinsmeister3, David a. Katzka2 1 Internal Medicine, Mayo Clinic, Rochester, MN; 2Gastroenterology & Hepatology, Mayo Clinic, Rochester, MN; 3Biomedical Statistics & Informatics, Mayo Clinic, Rochester, MN Background: Current screening guidelines for Barrett’s esophagus (BE) recommend patients undergo repeat esophagogastroduodenoscopy (EGD) after the diagnosis of Barrett’s to exclude prevalent dysplasia that could potentially alter management and surveillance protocols. The yield of repeat EGD in this setting remains unknown. Aim: Using a large population cohort of BE patients we aimed to assess the frequency with which repeat EGD revealed high-grade dysplasia (HGD) or esophageal adenocarcinoma (EAC) not detected during index endoscopy and to identify potential risk factors for finding dysplasia. Methods: A population based cohort of BE cases was created using resources of the Rochester Epidemiology Project in Olmsted County, Minnesota. Demographic, clinical and endoscopic data was abstracted from medical records. Cases were identified from 1976 to 2011. The rate of HGD or EAC on repeat EGD in patients with initially nondysplastic or low-grade dysplasia BE was obtained. Those not having a repeat EGD within 2 years of diagnosis were excluded. Results: 494 patients with newly diagnosed BE were identified. 445 had no dysplasia or LGD on index EGD, of which 178 (40%) underwent repeat EGD within 2 years. For this group, mean age (SD) was 61 (13), 120 (67%) were males and 160 (90%) were Caucasian. Four (2.2%) had higher grades of dysplasia on repeat EGD (2 unifocal HGDs after 39 and 105 days; 2 EACs after 239 and 397 days). All 4 were Caucasian males with a mean age of 69 (range 65-73), three of which had long segment Barrett’s esophagus (range 10-13 cm) with an inadequate number of biopsies (range 4-12 pieces) during index EGD. Both EACs arose from a single nodule or area of induration not previously visualized. They were managed with esophagogastrectomy or
www.giejournal.org
Abstracts
endoscopic mucosal resection and both patients are living (follow up of 9 and 6 years, respectively). Both patients with HGD elected to observe. One regressed to no dysplasia (follow up of 9 years). The other progressed to carcinoma after 3 years, requiring esophagogastrectomy, and died after a total 17 years of follow up due to other causes. The 441 patients who were not found to have HGD or EAC within 2 years had a median follow up time (range) of 7.8 years (29 years), during which 15 progressed to HGD or EAC. Conclusions: 1. The miss rate of HGD and EAC during the initial diagnosis of BE (2.2%) is not insignificant. 2. Older Caucasian males with long segment Barrett’s may have a higher miss rate. 3. Initial screening and follow up of Barrett’s patients is not performed in a majority of patients. 4. Given the high rate of prevalent than incident cancer in Barrett’s follow up, we should consider placement of greater resources into the initial screening and immediate follow up of Barrett’s patients.
163 Modifiable Risk Factors Predict Recurrence of Barrett’s Esophagus After Successful Radiofrequency Ablation W. Asher Wolf*, Cary C. Cotton, Nan LI, Sarina Pasricha, Kelly E. Hathorn, Melissa Spacek, Susan E. Moist, Ryan D. Madanick, Evan S. Dellon, Nicholas J. Shaheen Center for Esophageal Diseases and Swallowing, University of North Carolina School of Medicine, Chapel Hill, NC Background: Radiofrequency ablation (RFA) is a safe and effective treatment for Barrett’s Esophagus (BE), but intestinal metaplasia (IM) sometimes recurs in the tubular esophagus after successful treatment. We assessed predictors of recurrence after complete eradication of intestinal metaplasia (CEIM) in patients with BE treated with RFA. Methods: We conducted a retrospective study of adult patients who underwent RFA for BE at the University of North Carolina Hospitals between March 16, 2006 and August 30, 2013. We assessed multiple potential predictors of recurrence of IM including age, race, sex, BMI, alcohol and tobacco history, measures of BE burden, hiatal hernia and other variables (table). The first surveillance endoscopy (EGD) was defined as the first EGD without endoscopic or histologic evidence of IM; the second surveillance EGD was the first at risk for recurrent BE. Surveillance endoscopy utilized systematic 4 quadrant biopsies at 1 cm intervals throughout the previous BE segment, with targeted biopsies of any mucosal abnormalities. Recurrence was defined as any histology demonstrating IM in the tubular esophagus during endoscopic surveillance, regardless of endoscopic appearance. Bivariate analyses were performed using SAS 9.3. A multivariate logistic regression model was constructed including any variables with p!0.2 and reduced until all variables had p!0.05. Results: Of 291 patients who underwent RFA for BE, 229 (79%) achieved CEIM, and 180 (60%) underwent at least one additional surveillance endoscopy. Of those, 20 (11%) had recurrence of any BE: 10 (50%) with non-dysplastic BE (NDBE), 2 (10%) with low grade dysplasia (LGD), 4 (20%) with high grade dysplasia (HGD), 2 (10%) with intramucosal esophageal adenocarcinoma (IMC), and 2 (10%) with EAC. Six patients (3%) experienced disease progression: 2 from LGD to HGD, 2 from HGD to IMC, and 2 from HGD to EAC. Compared to those without recurrence, those with recurrence were younger (63 vs 69, pZ0.01), less likely to drink alcohol (20 vs 48%, pZ0.02), and more likely to use tobacco (40 vs 12%, pZ0.0009). On multivariate logistic regression after adjusting for age, Prague M length, and histology, the adjusted odds ratio (aOR) for recurrence with current tobacco use was 4.76 (95% confidence interval (CI) 1.49-15.22) while alcohol use was protective (aOR 0.30, 95% CI 0.09-0.99). Compared to a non-smoking drinker, the aOR for recurrence in a smoking non-drinker was 17 [95% CI 2.7-108.3]. Conclusion: Modifiable exposures are associated with recurrence after successful eradication of BE with RFA. While it is premature to recommend alcohol consumption to patients with BE, tobacco abstinence should be recommended. This work joins other recent data suggesting that alcohol consumption may be protective in the development or recurrence of BE.
Table 1. Characteristics of Patients with Progression and Those
Without
Age mean years SD White n (%) Male n (%) BMI mean SD Current Alcohol n (%) Current Tobacco n (%) Current NSAID n (%) Anti-Reflux Surgery n (%) Hiatal Hernia n (%) Prague C mean cm SD
www.giejournal.org
All (n[180)
Recurrence (n[20)
No Recurrence (n[160)
pvalue
68 10 171 (98) 132 (73) 30 5.7 81 (45) 27 (15) 86 (48) 8 (4) 161 (89) 1.2 2.3
63 10 20 (100) 13 (65) 31 8.5 4 (20) 8 (40) 6 (30) 0 (0) 18 (90) 1.7 2.8
69 10 151 (98) 119 (74) 29 5.3 77 (48) 19 (12) 80 (50) 8 (5) 143 (89) 1.1 2.2
0.01 0.52 0.37 0.62 0.02 0.0009 0.09 0.31 0.93 0.36
Prague M mean cm SD Days Treatment mean SD Days Surveillance mean SD Worst Treatment Histology NDBE LGD HGD EAC, non-invasive
All (n[180)
Recurrence (n[20)
No Recurrence (n[160)
pvalue
4.4 3.0 171 166 863 520
5.5 2.9 170 131 1032 608
4.2 2.9 171 170 842 506
0.051 0.58 0.17
5 (3) 56 (31) 102 (57) 17 (9)
1 (5) 6 (30) 12 (60) 1 (5)
0.82 4 50 90 16
(3) (31) (56) (10)
Table 2. Odds Ratios from Logistic Regression Predicting Recurrence Current Alcohol Current Tobacco
Unadjusted OR [95% CI]
Adjusted* OR [95% CI]
0.26 [0.08, 0.85] 5.02 [1.76, 14.30]
0.30 [0.09, 0.99] 4.76 [1.49, 15.22]
Adjusted for age, BE length (prague M), and treatment histology.
164 A FISH Biomarker Panel for the Prediction of High-grade Dysplasia and Adenocarcinoma in Non-Dysplastic Barrett’s Esophagus; Results From a Long-Term Prospective Cohort Study Margriet R. Timmer*1,2, Chiu Ting Lau1,2, Wilda Rosmolen1, Sybren L. Meijer3, Marcel G. Dijkgraaf4, Rosalie C. Mallant-Hent5, Anton H. Naber6, Arnoud H. Van Oijen7, Lubbertus C. Baak8, Pieter Scholten9, Clarisse BöHmer10, Paul Fockens1, Jacques J. Bergman1, Kausilia K. Krishnadath1,2 1 Division of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, Netherlands; 2Center for Experimental and Molecular Medicine, Academic Medical Center, Amsterdam, Netherlands; 3Department of Pathology, Academic Medical Center, Amsterdam, Netherlands; 4Clinical Research Unit, Academic Medical Center, Amsterdam, Netherlands; 5Department of Gastroenterology, Flevoziekenhuis, Almere, Netherlands; 6Department of Gastroenterology, Tergooiziekenhuizen, Hilversum, Netherlands; 7Department of Gastroenterology, Medisch Centrum Alkmaar, Alkmaar, Netherlands; 8 Department of Gastroenterology, Onze Lieve Vrouwe Gasthuis, Amsterdam, Netherlands; 9Department of Gastroenterology, Sint Lucas Andreas Ziekenhuis, Amsterdam, Netherlands; 10Department of Gastroenterology, Spaarneziekenhuis, Hoofddorp, Netherlands Background: Barrett’s esophagus (BE) with no dysplasia is associated with a risk as low as 0.1% per year of developing esophageal adenocarcinoma (EAC). Predictive biomarkers would be of great clinical value in facilitating more cost-effective surveillance. Due to the low progression rate of non-dysplastic BE, biomarker studies require long-term follow-up as well as a robust sample size and are therefore scarce. Methods: We conducted a prospective multicenter cohort study in a community-based setting of Barrett patients with no dysplasia to determine genetic predictors of progression. All patients were enrolled in an endoscopic surveillance program. Genetic abnormalities were detected on endoscopic cytology brushes by fluorescence in situ hybridization (FISH) using a probe-set including probes for P16, P53, Her-2/neu, 20q, and MYC, and the chromosomal centromeric probes 7 and 17 to detect aneuploidy. All markers were dichotomized into normal/abnormal based on cut-off values determined using ROC curves. Endpoints were progression to high-grade dysplasia (HGD) or EAC. Results: A total of 428 patients were included in the study (345 males; age 59 12 y.o.; BE length 3 cm, IQR 2-6) Median follow-up was 45 months (IQR 35 - 72). There were 22 patients (5%) with histologic progression after review by 2 expert pathologists; 13 cases of HGD and 9 cases of EAC. Univariable analysis revealed that P16, and aneuploidy were significantly associated with progression, as well as the clinical variables age and maximum Barrett segment length (M). The remaining markers showed a nonsignificant tendency towards increased odds of progression. Patients who tested positive for P16 and/or aneuploidy were designated as marker-positive. Kaplan-Meier analysis revealed that there was a significant difference in time to progression between the two groups (Log-rank PZ0.009; Figure 1). The overall rate of progression to HGD/ EAC was 1.09% per patient-year. Marker-positive patients had a higher annual risk to progress to HGD/EAC (1.85%) than marker-negative patients (0.58%) (PZ0.015). In a multivariate proportional hazards model, controlling for M and age, a positive FISH result was a significant predictor of histological progression to HGD/EAC (HR 3.23; 95% CI 1.32-7.95). Conclusion: A FISH panel assessing aneuploidy and P16, can be used as a decision making tool to stratify non-dysplastic Barrett patients into low- and high-risk disease categories to improve the efficacy of surveillance programs.
Volume 79, No. 5S : 2014 GASTROINTESTINAL ENDOSCOPY AB115
identified from a group of subjects who previously completed a validated symptom questionnaire. Patients were randomized (stratified by age, sex and reflux symptoms) and invited to undergo either uTNE in a mobile research van (muTNE), uTNE in outpatient endoscopy suite (huTNE) or sEGD. uTNE was performed using a portable esophagoscope with a disposable sheath. Procedure performance characteristics (completion rate, acquisition of biopsies, duration) were systematically recorded. Validated tolerability scales (0 Z none and 10 Z severe) and adverse events questionnaires were administered 1 and 30 days after procedures. Results: 459 subjects were contacted and 209 (46%) agreed to undergo study procedures (muTNE nZ76, huTNE nZ72, sEGD nZ61). Baseline characteristics were comparable among the three groups (table 1). Participation rates were numerically higher in the unsedated arms (muTNE 47.5%, huTNE 45.7%) than in the sEGD arm (40.7%) (pZ0.27). Patients with acid reflux symptoms R1/week were more likely to participate (odds ratio 2.94, 95% confidence interval 1.47, 5.88). Complete evaluation of the esophagus was comparable using muTNE (99%), huTNE (96%) and sEGD (100%) techniques. However, successful biopsy acquisition was lower in the muTNE (79%) and huTNE (83%) groups compared to sEGD (100%) (pZ0.001). Mean duration (minutes) of esophageal examination was shorter in the sEGD arm (4.7) than in the muTNE (8.0) and huTNE (8.5) groups (p!0.001). However, recovery time was much longer for sEGD (67.3) compared to muTNE (15.5) and huTNE (18.5) techniques (p!0.001). While overall tolerability for sEGD was better than muTNE and huTNE (mean score 0.4 vs. 1.9 and 2.2 respectively, p!0.001), the majority of patients who underwent muTNE and huTNE were willing to undergo the same procedure again in future (79% and 83%, respectively). No serious adverse events were reported. 16 subjects (7.6%) were diagnosed with BE (table 2). Conclusion: In this first large randomized trial evaluating novel approaches for community screening for BE, unsedated mobile van and clinic screening with TNE was feasible and effective. The patients’ visit was significantly shorter with adequate tolerability, acceptability and safety profiles. Mobile and outpatient techniques may provide a cost-effective alternative to sEGD for BE screening. Table 1. Baseline characteristics of subjects participating in the three
arms of the randomized trial (total n[209). Data presented as number (percentage) or mean (+/- standard deviation). PPI [ proton pump inhibitors. Variable Age (year) Male gender White ethnicity BMI Waist to hip ratio Education Less than high school High school or some college College/professional training Employment Employed Unemployed/homemaker Retired Marital status Married Not married Frequency of reflux symptoms R1/week !1/week None Current use PPI drugs Current aspirin use Symptom somatisation score Charlston comorbidity index
muTNE (n[76)
huTNE (n[72)
sEGD (n[61)
65 (9) 32 (42.1%) 73 (96.0%) 29.0 (5.6) 0.91 (0.09)
64 (9) 35 (48.6%) 72 (100%) 30.5 (13.9) 0.92 (0.11)
66 (9) 29 (47.5%) 61 (100%) 28.8 (5.8) 0.91 (0.09)
2 (2.7%) 37 (49.3%) 36 (48.0%)
1 (1.4%) 40 (55.6%) 31 (43.1%)
1 (1.7%) 33 (55.0%) 26 (43.3%)
23 (30.7%) 6 (8.0%) 46 (61.3%)
29 (40.3%) 6 (8.3%) 37 (51.4%)
17 (28.3%) 5 (8.3%) 38 (63.3%)
57 (76.0%) 18 (24.0%)
62 (86.1%) 10 (13.9%)
42 (70.0%) 18 (30.0%)
11 (14.5%) 26 (34.2%) 38 (50%) 15 (19.7%) 38 (50.0%) 0.8 (1.3) 0.9 (1.3)
10 (13.8%) 32 (44.4%) 27 (37.5%) 9 (12.5%) 37 (51.4%) 0.9 (1.4) 0.9 (1.4)
8 (13.1%) 28 (45.9%) 25 (41.0%) 11 (18.0%) 36 (59.0%) 0.8 (0.5) 1.3 (2.0)
Table 2. Findings suggestive of reflux-related mucosal injury. Data
presented as number of patients (percentage) Findings Normal Hiatus hernia Suspected BE on endoscopy Confirmed BE on histology Erosive esophagitis
muTNE (n[76) 58 48 12 7 12
(76.3%) (63.2%) (15.8%) (9.2%) (15.8%)
huTNE (n[72) 36 41 11 3 25
(50.0%) (56.9%) (15.3%) (4.2%) (34.7%)
sEGD (n[61) 23 34 15 6 24
(37.7%) (55.7%) (24.6%) (9.8%) (39.3%)
P value !0.001 0.67 0.37 0.29 0.003
AB114 GASTROINTESTINAL ENDOSCOPY Volume 79, No. 5S : 2014
161 Are the Gastric Folds Still the Optimal Landmark for Defining the Distal Border of a Barrett’s Esophagus in a Western Population? Dirk SchöLvinck*1,2, Osamu Goto3, Cees a. Seldenrijk4, Jacques J. Bergman2, Naohisa Yahagi3, Bas L. Weusten1,2 1 Gastroenterology and Hepatology, St. Antonius Hospital, Nieuwegein, Netherlands; 2Gastroenterology & Hepatology, Academic Medical Center, Amsterdam, Netherlands; 3Cancer Center, Keio University, School of Medicine, Tokyo, Japan; 4Pathology, St. Antonius Hospital, Nieuwegein, Netherlands Background: The definition of the esophagogastric junction (EGJ) for a diagnosis of Barrett’s esophagus (BE) is not uniform: according to Western guidelines, the EGJ is located at the top of the gastric folds (GF), assessed during expiration. In contrast, in Japan the EGJ is defined by the distal end of the palisade vessels (PV). However, PV are faint and often not clearly visible in Western patients. Narrow Band Imaging (NBI) enhances contrast in mucosal structures and vessels. The new Olympus EXERA III endoscopy system combines bright NBI with dual focal high-resolution endoscopy, and may more easily visualize PV, even in Western BE patients. Aim: To evaluate the detection rate of PV using the new EXERA III system, to quantify the discordance between the Western and Japanese criteria for the distal border of BE, and to evaluate the presence of intestinal metaplasia (IM) in this "zone of discordance" (ZoD), in Western BE patients. Methods: Consecutive Dutch patients with BE segment of R 2 cm (no dysplasia or low-grade dysplasia) on previous endoscopy were enrolled. Endoscopies were performed by a Western expert endoscopist, assisted by a Japanese endoscopist. The distance of the GF and distal ends of PV (when visible) to the incisors were determined in in- and desufflated conditions. Any difference in distance between the top of the GF (assessed during expiration) and the distal end of the PV was considered a ZoD. Two biopsies were taken from the cardia (either 1 cm below the distal end of the PV, or 1 cm below the GF in case of absence of PV), two biopsies from the ZoD when present, and 4-quadrant biopsies starting 1 cm proximal of the GF to determine the presence of IM. Results: 25 patients (19 males, median 66 yrs, median BE C5M7) were included. PV were seen in 96%. They were more often visible during inspiration (96%) than in expiration (76%; pZ0.06). PV were located in a constant position during in- and expiration. In patients with visible PV, in 63% (nZ15) a ZoD was present during expiration with the GF median 1 cm (range 1-2) proximal to the PV. In patients with a ZoD, IM was present in 6/30 (20%) of the ZoD biopsies. This was more compared to the cardia biopsies (0/30; 0%; pZ0.01) and less compared to the biopsies obtained proximal of the GF (35/60; 58%; p! 0.001). In the 10 patients without ZoD biopsies of the cardia and proximal of the GF contained IM in 0% and 63% respectively. Conclusion: In Western patients PV are frequently seen with the latest version NBI-system. PV are best observed under optimal insufflation. In a substantial subset of patients the GF and PV differ in location and IM can be detected in biopsies from this "zone of discordance", whereas no IM is detected in biopsies distal to the PV. These results question the value of the GF as the distal margin of the Barrett’s segment.
162 The Yield of Early Follow-up Endoscopy in Patients With Initial Nondysplastic or Low-Grade Dysplasia Barrett’s Esophagus Kavel Visrodia*1, Prasad G. Iyer2, Cathy D. Schleck3, Alan R. Zinsmeister3, David a. Katzka2 1 Internal Medicine, Mayo Clinic, Rochester, MN; 2Gastroenterology & Hepatology, Mayo Clinic, Rochester, MN; 3Biomedical Statistics & Informatics, Mayo Clinic, Rochester, MN Background: Current screening guidelines for Barrett’s esophagus (BE) recommend patients undergo repeat esophagogastroduodenoscopy (EGD) after the diagnosis of Barrett’s to exclude prevalent dysplasia that could potentially alter management and surveillance protocols. The yield of repeat EGD in this setting remains unknown. Aim: Using a large population cohort of BE patients we aimed to assess the frequency with which repeat EGD revealed high-grade dysplasia (HGD) or esophageal adenocarcinoma (EAC) not detected during index endoscopy and to identify potential risk factors for finding dysplasia. Methods: A population based cohort of BE cases was created using resources of the Rochester Epidemiology Project in Olmsted County, Minnesota. Demographic, clinical and endoscopic data was abstracted from medical records. Cases were identified from 1976 to 2011. The rate of HGD or EAC on repeat EGD in patients with initially nondysplastic or low-grade dysplasia BE was obtained. Those not having a repeat EGD within 2 years of diagnosis were excluded. Results: 494 patients with newly diagnosed BE were identified. 445 had no dysplasia or LGD on index EGD, of which 178 (40%) underwent repeat EGD within 2 years. For this group, mean age (SD) was 61 (13), 120 (67%) were males and 160 (90%) were Caucasian. Four (2.2%) had higher grades of dysplasia on repeat EGD (2 unifocal HGDs after 39 and 105 days; 2 EACs after 239 and 397 days). All 4 were Caucasian males with a mean age of 69 (range 65-73), three of which had long segment Barrett’s esophagus (range 10-13 cm) with an inadequate number of biopsies (range 4-12 pieces) during index EGD. Both EACs arose from a single nodule or area of induration not previously visualized. They were managed with esophagogastrectomy or
www.giejournal.org
Abstracts
endoscopic mucosal resection and both patients are living (follow up of 9 and 6 years, respectively). Both patients with HGD elected to observe. One regressed to no dysplasia (follow up of 9 years). The other progressed to carcinoma after 3 years, requiring esophagogastrectomy, and died after a total 17 years of follow up due to other causes. The 441 patients who were not found to have HGD or EAC within 2 years had a median follow up time (range) of 7.8 years (29 years), during which 15 progressed to HGD or EAC. Conclusions: 1. The miss rate of HGD and EAC during the initial diagnosis of BE (2.2%) is not insignificant. 2. Older Caucasian males with long segment Barrett’s may have a higher miss rate. 3. Initial screening and follow up of Barrett’s patients is not performed in a majority of patients. 4. Given the high rate of prevalent than incident cancer in Barrett’s follow up, we should consider placement of greater resources into the initial screening and immediate follow up of Barrett’s patients.
163 Modifiable Risk Factors Predict Recurrence of Barrett’s Esophagus After Successful Radiofrequency Ablation W. Asher Wolf*, Cary C. Cotton, Nan LI, Sarina Pasricha, Kelly E. Hathorn, Melissa Spacek, Susan E. Moist, Ryan D. Madanick, Evan S. Dellon, Nicholas J. Shaheen Center for Esophageal Diseases and Swallowing, University of North Carolina School of Medicine, Chapel Hill, NC Background: Radiofrequency ablation (RFA) is a safe and effective treatment for Barrett’s Esophagus (BE), but intestinal metaplasia (IM) sometimes recurs in the tubular esophagus after successful treatment. We assessed predictors of recurrence after complete eradication of intestinal metaplasia (CEIM) in patients with BE treated with RFA. Methods: We conducted a retrospective study of adult patients who underwent RFA for BE at the University of North Carolina Hospitals between March 16, 2006 and August 30, 2013. We assessed multiple potential predictors of recurrence of IM including age, race, sex, BMI, alcohol and tobacco history, measures of BE burden, hiatal hernia and other variables (table). The first surveillance endoscopy (EGD) was defined as the first EGD without endoscopic or histologic evidence of IM; the second surveillance EGD was the first at risk for recurrent BE. Surveillance endoscopy utilized systematic 4 quadrant biopsies at 1 cm intervals throughout the previous BE segment, with targeted biopsies of any mucosal abnormalities. Recurrence was defined as any histology demonstrating IM in the tubular esophagus during endoscopic surveillance, regardless of endoscopic appearance. Bivariate analyses were performed using SAS 9.3. A multivariate logistic regression model was constructed including any variables with p!0.2 and reduced until all variables had p!0.05. Results: Of 291 patients who underwent RFA for BE, 229 (79%) achieved CEIM, and 180 (60%) underwent at least one additional surveillance endoscopy. Of those, 20 (11%) had recurrence of any BE: 10 (50%) with non-dysplastic BE (NDBE), 2 (10%) with low grade dysplasia (LGD), 4 (20%) with high grade dysplasia (HGD), 2 (10%) with intramucosal esophageal adenocarcinoma (IMC), and 2 (10%) with EAC. Six patients (3%) experienced disease progression: 2 from LGD to HGD, 2 from HGD to IMC, and 2 from HGD to EAC. Compared to those without recurrence, those with recurrence were younger (63 vs 69, pZ0.01), less likely to drink alcohol (20 vs 48%, pZ0.02), and more likely to use tobacco (40 vs 12%, pZ0.0009). On multivariate logistic regression after adjusting for age, Prague M length, and histology, the adjusted odds ratio (aOR) for recurrence with current tobacco use was 4.76 (95% confidence interval (CI) 1.49-15.22) while alcohol use was protective (aOR 0.30, 95% CI 0.09-0.99). Compared to a non-smoking drinker, the aOR for recurrence in a smoking non-drinker was 17 [95% CI 2.7-108.3]. Conclusion: Modifiable exposures are associated with recurrence after successful eradication of BE with RFA. While it is premature to recommend alcohol consumption to patients with BE, tobacco abstinence should be recommended. This work joins other recent data suggesting that alcohol consumption may be protective in the development or recurrence of BE.
Table 1. Characteristics of Patients with Progression and Those
Without
Age mean years SD White n (%) Male n (%) BMI mean SD Current Alcohol n (%) Current Tobacco n (%) Current NSAID n (%) Anti-Reflux Surgery n (%) Hiatal Hernia n (%) Prague C mean cm SD
www.giejournal.org
All (n[180)
Recurrence (n[20)
No Recurrence (n[160)
pvalue
68 10 171 (98) 132 (73) 30 5.7 81 (45) 27 (15) 86 (48) 8 (4) 161 (89) 1.2 2.3
63 10 20 (100) 13 (65) 31 8.5 4 (20) 8 (40) 6 (30) 0 (0) 18 (90) 1.7 2.8
69 10 151 (98) 119 (74) 29 5.3 77 (48) 19 (12) 80 (50) 8 (5) 143 (89) 1.1 2.2
0.01 0.52 0.37 0.62 0.02 0.0009 0.09 0.31 0.93 0.36
Prague M mean cm SD Days Treatment mean SD Days Surveillance mean SD Worst Treatment Histology NDBE LGD HGD EAC, non-invasive
All (n[180)
Recurrence (n[20)
No Recurrence (n[160)
pvalue
4.4 3.0 171 166 863 520
5.5 2.9 170 131 1032 608
4.2 2.9 171 170 842 506
0.051 0.58 0.17
5 (3) 56 (31) 102 (57) 17 (9)
1 (5) 6 (30) 12 (60) 1 (5)
0.82 4 50 90 16
(3) (31) (56) (10)
Table 2. Odds Ratios from Logistic Regression Predicting Recurrence Current Alcohol Current Tobacco
Unadjusted OR [95% CI]
Adjusted* OR [95% CI]
0.26 [0.08, 0.85] 5.02 [1.76, 14.30]
0.30 [0.09, 0.99] 4.76 [1.49, 15.22]
Adjusted for age, BE length (prague M), and treatment histology.
164 A FISH Biomarker Panel for the Prediction of High-grade Dysplasia and Adenocarcinoma in Non-Dysplastic Barrett’s Esophagus; Results From a Long-Term Prospective Cohort Study Margriet R. Timmer*1,2, Chiu Ting Lau1,2, Wilda Rosmolen1, Sybren L. Meijer3, Marcel G. Dijkgraaf4, Rosalie C. Mallant-Hent5, Anton H. Naber6, Arnoud H. Van Oijen7, Lubbertus C. Baak8, Pieter Scholten9, Clarisse BöHmer10, Paul Fockens1, Jacques J. Bergman1, Kausilia K. Krishnadath1,2 1 Division of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, Netherlands; 2Center for Experimental and Molecular Medicine, Academic Medical Center, Amsterdam, Netherlands; 3Department of Pathology, Academic Medical Center, Amsterdam, Netherlands; 4Clinical Research Unit, Academic Medical Center, Amsterdam, Netherlands; 5Department of Gastroenterology, Flevoziekenhuis, Almere, Netherlands; 6Department of Gastroenterology, Tergooiziekenhuizen, Hilversum, Netherlands; 7Department of Gastroenterology, Medisch Centrum Alkmaar, Alkmaar, Netherlands; 8 Department of Gastroenterology, Onze Lieve Vrouwe Gasthuis, Amsterdam, Netherlands; 9Department of Gastroenterology, Sint Lucas Andreas Ziekenhuis, Amsterdam, Netherlands; 10Department of Gastroenterology, Spaarneziekenhuis, Hoofddorp, Netherlands Background: Barrett’s esophagus (BE) with no dysplasia is associated with a risk as low as 0.1% per year of developing esophageal adenocarcinoma (EAC). Predictive biomarkers would be of great clinical value in facilitating more cost-effective surveillance. Due to the low progression rate of non-dysplastic BE, biomarker studies require long-term follow-up as well as a robust sample size and are therefore scarce. Methods: We conducted a prospective multicenter cohort study in a community-based setting of Barrett patients with no dysplasia to determine genetic predictors of progression. All patients were enrolled in an endoscopic surveillance program. Genetic abnormalities were detected on endoscopic cytology brushes by fluorescence in situ hybridization (FISH) using a probe-set including probes for P16, P53, Her-2/neu, 20q, and MYC, and the chromosomal centromeric probes 7 and 17 to detect aneuploidy. All markers were dichotomized into normal/abnormal based on cut-off values determined using ROC curves. Endpoints were progression to high-grade dysplasia (HGD) or EAC. Results: A total of 428 patients were included in the study (345 males; age 59 12 y.o.; BE length 3 cm, IQR 2-6) Median follow-up was 45 months (IQR 35 - 72). There were 22 patients (5%) with histologic progression after review by 2 expert pathologists; 13 cases of HGD and 9 cases of EAC. Univariable analysis revealed that P16, and aneuploidy were significantly associated with progression, as well as the clinical variables age and maximum Barrett segment length (M). The remaining markers showed a nonsignificant tendency towards increased odds of progression. Patients who tested positive for P16 and/or aneuploidy were designated as marker-positive. Kaplan-Meier analysis revealed that there was a significant difference in time to progression between the two groups (Log-rank PZ0.009; Figure 1). The overall rate of progression to HGD/ EAC was 1.09% per patient-year. Marker-positive patients had a higher annual risk to progress to HGD/EAC (1.85%) than marker-negative patients (0.58%) (PZ0.015). In a multivariate proportional hazards model, controlling for M and age, a positive FISH result was a significant predictor of histological progression to HGD/EAC (HR 3.23; 95% CI 1.32-7.95). Conclusion: A FISH panel assessing aneuploidy and P16, can be used as a decision making tool to stratify non-dysplastic Barrett patients into low- and high-risk disease categories to improve the efficacy of surveillance programs.
Volume 79, No. 5S : 2014 GASTROINTESTINAL ENDOSCOPY AB115