- Email: [email protected]
536 Diagnostic Accuracy of Probe-Based Confocal Laser Endomicroscopy (Pcle) Compared to Random Biopsies in Patients Undergoing Endoscopic Surveillance of Barrett's Esophagus
Abstracts
Backgrounds & Aims: Although radical surgery is recommended for patients not meeting the curative criteria for endoscopic submucosal dissection (ESD) of early gastric cancer (EGC) in the European and Japanese guidelines due to the potential risk of lymph node metastasis (LNM), this recommendation may be excessive because LNM is found in only 5–10% of them. Thus, based on the analysis of risk factors for LNM of EGC, we established a seven-point risk-scoring system (3 points for lymphatic invasion, 1 point each for tumor size of >30 mm, deep submucosal invasion, venous invasion, and positive vertical margin) with three risk stratification, named eCura system, that predict LNM after ESD (low (0–1 point: 2.5% risk), intermediate (2–4 points: 6.7%), and high (5–7 points: 22.7%)). However, it has been still unclear whether this system contribute to the selection of patients who need radical surgery after ESD. Thus, in this study, we aimed to clarify the utility of eCura system for deciding treatment strategy after ESD by a comparative study. Methods: Of 15,785 patients, 1,969 not meeting the curative criteria for ESD of EGC were included in this multicenter study at 19 institutions between 2000 and 2011. Based on the treatment strategy after ESD, we had the radical surgery (n Z1,064) and follow-up (no additional treatment, n Z905) groups. After eCura system was applied to these patients, cancer-specific mortality and cancer recurrence at each risk category in eCura system was compared between these two groups. In this study, recurrence was defined as tumor relapse in the lymph nodes and/or other organs after a series of treatments for EGC. Results: With a median follow-up period of 65 months, cancer-specific survival in the follow-up group was significantly lower than that in the radical surgery group in patients at high- (90.1% and 96.1% in 5-year, respectively, p Z0.001) and intermediate-risk (96.0% and 99.2% in 5-year, respectively, p Z0.007), whereas that was not significantly different between the two groups in patients at low-risk (99.6% and 99.7%, respectively, p Z0.46). Cox analysis, adjusted by age, gender, location, and so on, revealed that, in patients at high-risk, cancer-specific mortality in the follow-up group tended to be higher than that in the radical surgery group [hazard ratio (95% confidence interval) Z 2.66 (0.95–7.48), p Z0.063], and that cancer recurrence in the former was significantly higher [3.13 (1.16–8.46), p Z0.024], whereas cancer-specific mortality and cancer recurrence showed no significant difference between two groups in those at intermediate- or low-risk. Conclusions: This study demonstrated that radical surgery is recommended for patients at high-risk in eCura system after ESD not meeting the curative criteria for EGC, whereas follow-up with no additional treatment might be an acceptable option for those at low-risk. Cox proportional-hazards model for risk of cancer-specific and all-cause mortalities in patients who were followed up without additional treatment after ESD compared with those who underwent radical surgery after ESD at each risk category.
Risk category Low
Intermediate
High
Treatment strategy after ESD Radical surgery Follow-up Radical surgery Follow-up Radical surgery Follow-up
Cancer-specific mortality HR
95% CI
1
Reference
p value
Cancer recurrence HR
95% CI
1
Reference
p value
1.44 0.12-17.2 1 Reference
0.78
0.74 0.10-5.63 1 Reference
0.78
1.66 0.43-6.40 1 Reference
0.46
2.00 0.54-7.36 1 Reference
0.30
2.66
0.063
3.13
0.024
0.95-7.48
1.16-8.46
ESD, endoscopic submucosal dissection; HR, hazard ratio; CI, confidence interval.
535 Incidence of Malignant Progression in Persistent Nondysplastic Barrett’s Esophagus, a Dutch Nationwide Cohort Study Yonne Peters*1, Judith Honing1, Wietske Kievit2, Iris D. Nagtegaal3, Peter D. Siersema1 1 Department of Gastroenterology and Hepatology, Radboud University Medical Center, Nijmegen, Netherlands; 2Radboud Institute for Health Sciences, Radboud University Medical Center, Nijmegen, Netherlands; 3 Department of Pathology, Radboud University Medical Center, Nijmegen, Netherlands Background: The risk of esophageal adenocarcinoma (EAC) in patients with nondysplastic Barrett’s esophagus (NDBE) may have been overestimated in some studies. Lower progression rates could lead to alteration of the current surveillance strategy, reducing its associated risks and costs. The aim of this study was to evaluate whether persistence of NDBE over consecutive surveillance endoscopies identifies patients at a very low risk for malignant progression. Methods: In this population-
AB76 GASTROINTESTINAL ENDOSCOPY Volume 85, No. 5S : 2017
based retrospective study, patients with a first diagnosis of NDBE between 2003 and 2012 were selected using PALGA, a nationwide registry of histopathology diagnoses in the Netherlands. Patients were followed until the development of EAC or highgrade dysplasia (HGD), or until the last endoscopy contact with biopsy sampling (through May 2016). Persistent NDBE was defined as NDBE at the index and the first follow-up endoscopy with at least one year of follow-up after the first surveillance endoscopy. Results: In this study 12,731 patients with NDBE were included, with a total follow-up of 64,898 (median 4.4 (IQR 3.0 – 6.8)) years. Median duration until first follow-up endoscopy was 2.2 years (IQR 1.5 – 3.2). During the study period, malignant progression was seen in 437 patients (3.4%), after a median follow-up of 5.1 years (IQR 3.2 – 7.3). This resulted in a progression rate to EAC of 0.47 (95% CI: 0.42 – 0.53) per 100 person-years and a progression rate to HGD/EAC combined of 0.68 (95% CI: 0.62 – 0.74) per 100 person-years. In the total study cohort, 11,854 patients (93%) were not found to have developed dysplasia at first follow-up endoscopy, whereas 675 (5.3%) and 202 (1.6%) patients showed progression to low-grade dysplasia and HGD/EAC, respectively. In patients with two consecutive endoscopies showing NDBE progression rates to EAC alone and HGD/EAC combined were 0.26 (95% CI: 0.22 – 0.32) and 0.40 (95% CI: 0.35 – 0.47) per 100 person-years, respectively. The risk of EAC in patients with persistent NDBE decreased to 0.19 (95% CI: 0.12 – 0.31) in patients with an interval of at least 3 years between index NDBE diagnosis and first follow-up endoscopy. Patients with five or more consecutive non-progressive endoscopies had a 75% relative risk reduction for development of EAC (progression rate: 0.13 (95% CI: 0.11- 0.41) per 100 person-years). Conclusion: Persistent NDBE at two consecutive surveillance endoscopies reduces the risk of malignant progression towards EAC by almost twofold (0.47 to 0.26 per 100 person-years), which suggests that prolonging the surveillance intervals in patients with persistent NDBE could be considered. As progression risk reduces even further after more consecutive negative follow-up endoscopies, future studies should aim to establish the group of patients in whom surveillance could safely be discontinued.
536 Diagnostic Accuracy of Probe-Based Confocal Laser Endomicroscopy (Pcle) Compared to Random Biopsies in Patients Undergoing Endoscopic Surveillance of Barrett’s Esophagus Tilak U. Shah*1,2, Robert Lippman1,2, Divyanshoo R. Kohli2,1, Pritesh R. Mutha1,2, Sanjeev S. Solomon2,1, Alvin M. Zfass1,2 1 Gastroenterology, Hunter Holmes Mcguire VA Medical Center, Richmond, VA; 2Gastroenterology, Virginia Commonwealth University, Richmond, VA Background: In patients undergoing endoscopic surveillance of Barrett’s esophagus, the current standard of random 4-quadrant biopsies obtained at 1-2 cm intervals (Seattle protocol) misses 10-50% of esophageal neoplasms, may increase risk of bleeding, and may not be cost-effective. Probe based confocal laser endomicroscopy (pCLE) permits real-time in-vivo histologic assessment of esophageal mucosa at the time of upper endoscopy. Several guidelines have endorsed the technology as an alternative to random biopsies. These guidelines are based on randomized studies that compared pCLE to white light endoscopy, or used an endoscope-based version (eCLE) that is not commercially available. Prospective studies comparing accuracy of pCLE to Seattle protocol biopsies in clinical practice are lacking. Methods: Consecutive patients referred for surveillance endoscopy for Barrett’s esophagus underwent high-definition white light and narrow band imaging followed by pCLE and targeted biopsy or mucosal resection. Subsequently, Seattle protocol biopsies were obtained during the same session. Baseline variables, real-time pCLE interpretation, and histology results were prospectively recorded. An expert endomicroscopist performed blinded review of pCLE sequences, and an expert pathologist performed blinded review of histologic specimens. A sample size of 64 patients was calculated a priori based on 3% estimated prevalence of high-grade dysplasia (HGD) or cancer, 80% per-patient pCLE specificity for HGD or cancer, power 80%, and alpha 0.05. Results: 66 patients were included in the study (mean age 63 years, 98% male, and 92% white). Mean Barrett’s segment length was C2M3. The median time to perform pCLE was 7 minutes. The prevalence of HGD or cancer was 4.47% (2 cancers, 1 HGD). Both real-time and blinded pCLE correctly identified all cases of cancer. For the primary outcome, accuracy of real-time pCLE for diagnosing HGD/cancer compared to non-blinded pathologist interpretation was as follows: sensitivity 66.7%, specificity 96.8%, negative predictive value 50%, and positive predictive value 98.4% (Table 1). For HGD and cancer, inter-observer agreement was substantial between realtime and blinded endomicroscopists (kappaZ0.6), and was perfect between blinded and non-blinded pathologists (kappaZ1). For low-grade dysplasia interobserver agreement was minimal between endomicroscopists as well as between pathologists (kappaZ0.2 in both groups; Table 2). Conclusions: pCLE demonstrates high specificity for detecting dysplasia, but lower sensitivity may limit its widespread adoption as a replacement to random biopsies. There is substantial disagreement between endomicroscopists and between pathologists for the
www.giejournal.org
Abstracts
diagnosis of low-grade dysplasia. Cost-effectiveness analyses may help identify the optimal role for this technology in patients with Barrett’s esophagus.
Table 1: Accuracy of probe-based confocal endomicroscopy (pCLE) compared to histology for high-grade dysplasia or cancer
Real-time pCLE interpretation
Blinded pCLE interpretation
Non-blinded pathologist interpretation
Blinded pathologist interpretation
Sensitivity 66.7% Specificity 96.8% NPV 50% PPV 98.4% Sensitivity 66.7% Specificity 93.7% NPV 33.3% PPV 98.3%
Sensitivity 66.7% Specificity 96.8% NPV 50% PPV 98.4% Sensitivity 66.7% Specificity 93.7% NPV 33.3% PPV 98.3%
PPV – positive predictive value; NPV – negative predictive value.
Table 2: Accuracy of probe-based confocal endomicroscopy (pCLE) compared to histology for low-grade dysplasia
Real-time pCLE interpretation
Blinded pCLE interpretation
Non-blinded pathologist interpretation
Blinded pathologist interpretation
Sensitivity 60% Specificity 86.9% NPV 27.3% PPV 96.4% Sensitivity 0% Specificity 93.4% NPV 0% PPV 91.9%
Sensitivity 31.6% Specificity 89.4% NPV 54.5% PPV 76.4% Sensitivity 10.5% Specificity 95.7% NPV 50% PPV 72.6%
PPV – positive predictive value; NPV – negative predictive value.
537 Quality of Endoscopic Surveillance of Barrett’s Esophagus in Denmark Jes S. Vogt*1, Anders Christian Larsen1, Thorbjørn Sommer2, Per Ejstrud1 1 Department of Gastrointestinal Surgery, Aalborg University Hospital, Aalborg, Denmark; 2Department of Surgery, Region Hospital Randers, Randers, Denmark Background: The increasing incidence of esophageal adenocarcinoma and novel endoluminal treatments of Barrett’s neoplasia, emphasizes the importance of high quality surveillance of Barrett’s esophagus (BE). Adherence to surveillance guidelines in Denmark has never been investigated. The aim of this study was to perform a retrospective national study evaluating the compliance with Danish national guidelines. Methods: During a three year period, a total of 3692 patients with BE were identified using The Danish Pathology Registry. 300 patients were included by drawing a simple random sample. Information regarding description of the BE segment, biopsy protocol, communication with the pathologist and planned followup endoscopy, was obtained. Results: From the subset of 300 patients, we excluded 31 patients (10%) due to missing reports. 83 patients (28%) had no endoscopic evidence of BE according to the endoscopy reports and were also excluded. BE was macroscopically suspected by the endoscopist in 186 patients (62%) and these patients were included in this study. Prague C&M classification was recorded in 63 of the endoscopy reports (34%). The BE segment was stratified by lengths of 1-5 cm, 510 cm and 10-15 cm and the median number of biopsy specimens obtained in each segment was 4 (interquartile range (IQR), 3-6), 5.5 (IQR, 4-9) and 4 (IQR, 1-6). According to the modified Seattle biopsy protocol in the Danish guidelines, a correct minimum number of biopsies were obtained in only 12%, 8% and 0% of cases, respectively. The median number of biopsies in all patients with endoscopic evidence of BE was 4 (IQR 3-6). 95% of endoscopists reported in the pathology requisition that the indication for histological evaluation was BE. 28% described the exact location of the biopsies in the pathology requisition. Patients with BE and no dysplasia, had endoscopic surveillance performed after a median of 24 months (IQR, 6-24). Conclusion: Adherence to Danish guidelines is poor and may be associated with insufficient quality of BE surveillance and reduced detection of neoplasia in BE patients. The substantial lack of endoscopic evidence of BE and non-compliance to guidelines in our study, call for caution when interpreting and conducting studies based on pathology registry. Previously published studies based on The Danish Pathology Registry and the quality of endoscopic surveillance may substantially underestimate the incidence of adenocarcinoma in BE patients.
www.giejournal.org
538 Feasibility and Performance Characteristics of a Novel Disposable Transnasal Capsule Device for Barrett’s Esophagus Screening: A Prospective International Multicenter Trial Prachi Pophali1, Sami Sarmed2, Magnus Halland3, Massimiliano DiPietro3, Felicity Enders1, Jacobo Ortiz Fernández-Sordo2, Jonathan White2, Michele L. Johnson1, Indra Guha2, Rebecca C. Fitzgerald3, Krish Ragunath2, Prasad G. Iyer*1 1 Mayo Clinic, Rochester, MN; 2Queens Medical Centre, Nottingham, United Kingdom; 3University of Cambridge, Cambridge, United Kingdom Background: Screening for Barrett’s esophagus (BE) with standard endoscopy (SE) is expensive. Unsedated transnasal endoscopy (uTNE) is a feasible and safe alternative for BE screening. We aimed to assess the technical feasibility, safety and performance characteristics of clinic-based uTNE using a novel single use transnasal capsule based device (EG Scan II) compared to SE. Methods: A prospective study was conducted at three centers (2 United Kingdom, 1 United States). Patients referred for clinical SE with and without known BE were invited to participate. uTNE with the EG Scan device (consisting of a 6 mm diameter disposable capsule attached to a controller and air compressor) was performed first followed by SE. Both procedures were conducted by different endoscopists blinded to the indication and outcome of either procedure. Results of the SE were considered as the gold standard. Patient symptoms (gagging, choking, discomfort, nasal pain) were recorded on a 10-point visual analogue scale (VAS) (10 being the worst), in addition to overall preference and tolerability on a 10-point VAS (10 being the best). Results: 200 patients, 100 with and 100 without known BE, 33% females and mean (SD) age of 57.8 (14) years participated. Median (IQR) length of the BE segment was 2 (1, 4) cm. 178 (89%) patients completed both procedures. 22 (11%) did not complete the EG Scan exam (1 poor tolerability after intubation, 21 inability to traverse nasal passages). The median symptom scores for the EG Scan were excellent ranging from 0-2. Overall tolerability was good with a median VAS score of 8 for the EG Scan, compared to 7 for SE. 54.5% of the participants preferred the EG Scan, while 16.8% preferred SE. 1 patient developed self-limiting epistaxis following uTNE. Sensitivity and specificity of EG Scan for BE diagnosis was 89.4% (95% CI 83.3- 95.6) and 90.3% (95% CI 84 – 96.7) respectively. Sensitivity was superior for long segment BE (93%) than short segment BE (87%). Conclusions: uTNE with the EG Scan device is a safe, welltolerated and accurate alternative for BE screening. Thus, it could potentially serve as a community and office based tool for BE screening.
539 Dysplasia Recurrence After Complete Eradication of Intestinal Metaplasia (CEIM) and Dysplasia (CED) in Barrett’s Esophagus Treated With Endoscopic Therapy: Results From an International, Multi-Center Consortium Rajesh Krishnamoorthi*1, Sreekar Vennelaganti9, Alessandro Repici8, Vani J. Konda3, Irving Waxman3, Stefan Seewald4, Matthew W. Stier3, Rehan Haidry6, Daniel Buckles2, Neil Gupta5, Ajay Bansal2, Sharad Mathur2, Mojtaba S. Olyaee2, Michael J. Bourke7, Gary W. Falk10, Ramprasad Jegadeesan2, Julie Nguyen2, Prashanth Vennalaganti9, Anusha Vittal2, Kevin F. Kennedy9, Prateek Sharma2,9, Andrew S. Ross1 1 Digestive Disease Institute, Virginia Mason Medical Center, Seattle, WA; 2 University of Kansas, Kansas City, KS; 3University of Chicago, Chicago, IL; 4Hirsladen, Zurich, Switzerland; 5Loyola University, Maywood, IL; 6 University college London Hospital, London, United Kingdom; 7 University of Sydney, Westmead, New South Wales, Australia; 8Instituto Clinico Humanitas, Milan, Italy; 9Kansas city VA Medical Center, Kansas city, MO; 10University of Pennsylvania Perelman School of Medicine, Philadelphia, PA Background: CEIM is more difficult to achieve than CED during endoscopic therapy for dysplastic BE. Hence, it is important to define the goal of endoscopic therapy for each individual patient – complete elimination of intestinal metaplasia (CEIM) or dysplasia (CED). We aimed to compare the risk of recurrent dysplastic BE in patients who achieved CEIM and CED. Methods: We reviewed all BE subjects who underwent endoscopic therapy in a multicenter prospective registry. BE subjects with LGD, HGD and EAC who reached CEIM or CED were included in this analysis. Dysplastic recurrence in the CEIM and CED groups was defined as presence of recurrent BE endoscopically (columnar mucosa in distal esophagus) with histologic evidence of any dysplasia or EAC after achieving CEIM and CED respectively. Cox proportional hazard models were used to compare the risk of dysplastic recurrence between CEIM and CED groups. Results: 266 subjects with dysplastic BE who underwent endoscopic therapy were included in the study. 191 subjects achieved CEIM and 75 subjects achieved CED. The mean (SD) age was 64 (11.2) years.240 (90.2%) were males. The median follow up of subjects was 31.2 months. Overall, 52 subjects (19.5%) had recurrence of dysplasia; the rate of recurrence was 21% in the CEIM group and 16% in the in the CED group. Baseline characteristics of CEIM and CED groups are summarized in table 1. The risk of dysplastic recurrence in CEIM group
Volume 85, No. 5S : 2017 GASTROINTESTINAL ENDOSCOPY AB77
Backgrounds & Aims: Although radical surgery is recommended for patients not meeting the curative criteria for endoscopic submucosal dissection (ESD) of early gastric cancer (EGC) in the European and Japanese guidelines due to the potential risk of lymph node metastasis (LNM), this recommendation may be excessive because LNM is found in only 5–10% of them. Thus, based on the analysis of risk factors for LNM of EGC, we established a seven-point risk-scoring system (3 points for lymphatic invasion, 1 point each for tumor size of >30 mm, deep submucosal invasion, venous invasion, and positive vertical margin) with three risk stratification, named eCura system, that predict LNM after ESD (low (0–1 point: 2.5% risk), intermediate (2–4 points: 6.7%), and high (5–7 points: 22.7%)). However, it has been still unclear whether this system contribute to the selection of patients who need radical surgery after ESD. Thus, in this study, we aimed to clarify the utility of eCura system for deciding treatment strategy after ESD by a comparative study. Methods: Of 15,785 patients, 1,969 not meeting the curative criteria for ESD of EGC were included in this multicenter study at 19 institutions between 2000 and 2011. Based on the treatment strategy after ESD, we had the radical surgery (n Z1,064) and follow-up (no additional treatment, n Z905) groups. After eCura system was applied to these patients, cancer-specific mortality and cancer recurrence at each risk category in eCura system was compared between these two groups. In this study, recurrence was defined as tumor relapse in the lymph nodes and/or other organs after a series of treatments for EGC. Results: With a median follow-up period of 65 months, cancer-specific survival in the follow-up group was significantly lower than that in the radical surgery group in patients at high- (90.1% and 96.1% in 5-year, respectively, p Z0.001) and intermediate-risk (96.0% and 99.2% in 5-year, respectively, p Z0.007), whereas that was not significantly different between the two groups in patients at low-risk (99.6% and 99.7%, respectively, p Z0.46). Cox analysis, adjusted by age, gender, location, and so on, revealed that, in patients at high-risk, cancer-specific mortality in the follow-up group tended to be higher than that in the radical surgery group [hazard ratio (95% confidence interval) Z 2.66 (0.95–7.48), p Z0.063], and that cancer recurrence in the former was significantly higher [3.13 (1.16–8.46), p Z0.024], whereas cancer-specific mortality and cancer recurrence showed no significant difference between two groups in those at intermediate- or low-risk. Conclusions: This study demonstrated that radical surgery is recommended for patients at high-risk in eCura system after ESD not meeting the curative criteria for EGC, whereas follow-up with no additional treatment might be an acceptable option for those at low-risk. Cox proportional-hazards model for risk of cancer-specific and all-cause mortalities in patients who were followed up without additional treatment after ESD compared with those who underwent radical surgery after ESD at each risk category.
Risk category Low
Intermediate
High
Treatment strategy after ESD Radical surgery Follow-up Radical surgery Follow-up Radical surgery Follow-up
Cancer-specific mortality HR
95% CI
1
Reference
p value
Cancer recurrence HR
95% CI
1
Reference
p value
1.44 0.12-17.2 1 Reference
0.78
0.74 0.10-5.63 1 Reference
0.78
1.66 0.43-6.40 1 Reference
0.46
2.00 0.54-7.36 1 Reference
0.30
2.66
0.063
3.13
0.024
0.95-7.48
1.16-8.46
ESD, endoscopic submucosal dissection; HR, hazard ratio; CI, confidence interval.
535 Incidence of Malignant Progression in Persistent Nondysplastic Barrett’s Esophagus, a Dutch Nationwide Cohort Study Yonne Peters*1, Judith Honing1, Wietske Kievit2, Iris D. Nagtegaal3, Peter D. Siersema1 1 Department of Gastroenterology and Hepatology, Radboud University Medical Center, Nijmegen, Netherlands; 2Radboud Institute for Health Sciences, Radboud University Medical Center, Nijmegen, Netherlands; 3 Department of Pathology, Radboud University Medical Center, Nijmegen, Netherlands Background: The risk of esophageal adenocarcinoma (EAC) in patients with nondysplastic Barrett’s esophagus (NDBE) may have been overestimated in some studies. Lower progression rates could lead to alteration of the current surveillance strategy, reducing its associated risks and costs. The aim of this study was to evaluate whether persistence of NDBE over consecutive surveillance endoscopies identifies patients at a very low risk for malignant progression. Methods: In this population-
AB76 GASTROINTESTINAL ENDOSCOPY Volume 85, No. 5S : 2017
based retrospective study, patients with a first diagnosis of NDBE between 2003 and 2012 were selected using PALGA, a nationwide registry of histopathology diagnoses in the Netherlands. Patients were followed until the development of EAC or highgrade dysplasia (HGD), or until the last endoscopy contact with biopsy sampling (through May 2016). Persistent NDBE was defined as NDBE at the index and the first follow-up endoscopy with at least one year of follow-up after the first surveillance endoscopy. Results: In this study 12,731 patients with NDBE were included, with a total follow-up of 64,898 (median 4.4 (IQR 3.0 – 6.8)) years. Median duration until first follow-up endoscopy was 2.2 years (IQR 1.5 – 3.2). During the study period, malignant progression was seen in 437 patients (3.4%), after a median follow-up of 5.1 years (IQR 3.2 – 7.3). This resulted in a progression rate to EAC of 0.47 (95% CI: 0.42 – 0.53) per 100 person-years and a progression rate to HGD/EAC combined of 0.68 (95% CI: 0.62 – 0.74) per 100 person-years. In the total study cohort, 11,854 patients (93%) were not found to have developed dysplasia at first follow-up endoscopy, whereas 675 (5.3%) and 202 (1.6%) patients showed progression to low-grade dysplasia and HGD/EAC, respectively. In patients with two consecutive endoscopies showing NDBE progression rates to EAC alone and HGD/EAC combined were 0.26 (95% CI: 0.22 – 0.32) and 0.40 (95% CI: 0.35 – 0.47) per 100 person-years, respectively. The risk of EAC in patients with persistent NDBE decreased to 0.19 (95% CI: 0.12 – 0.31) in patients with an interval of at least 3 years between index NDBE diagnosis and first follow-up endoscopy. Patients with five or more consecutive non-progressive endoscopies had a 75% relative risk reduction for development of EAC (progression rate: 0.13 (95% CI: 0.11- 0.41) per 100 person-years). Conclusion: Persistent NDBE at two consecutive surveillance endoscopies reduces the risk of malignant progression towards EAC by almost twofold (0.47 to 0.26 per 100 person-years), which suggests that prolonging the surveillance intervals in patients with persistent NDBE could be considered. As progression risk reduces even further after more consecutive negative follow-up endoscopies, future studies should aim to establish the group of patients in whom surveillance could safely be discontinued.
536 Diagnostic Accuracy of Probe-Based Confocal Laser Endomicroscopy (Pcle) Compared to Random Biopsies in Patients Undergoing Endoscopic Surveillance of Barrett’s Esophagus Tilak U. Shah*1,2, Robert Lippman1,2, Divyanshoo R. Kohli2,1, Pritesh R. Mutha1,2, Sanjeev S. Solomon2,1, Alvin M. Zfass1,2 1 Gastroenterology, Hunter Holmes Mcguire VA Medical Center, Richmond, VA; 2Gastroenterology, Virginia Commonwealth University, Richmond, VA Background: In patients undergoing endoscopic surveillance of Barrett’s esophagus, the current standard of random 4-quadrant biopsies obtained at 1-2 cm intervals (Seattle protocol) misses 10-50% of esophageal neoplasms, may increase risk of bleeding, and may not be cost-effective. Probe based confocal laser endomicroscopy (pCLE) permits real-time in-vivo histologic assessment of esophageal mucosa at the time of upper endoscopy. Several guidelines have endorsed the technology as an alternative to random biopsies. These guidelines are based on randomized studies that compared pCLE to white light endoscopy, or used an endoscope-based version (eCLE) that is not commercially available. Prospective studies comparing accuracy of pCLE to Seattle protocol biopsies in clinical practice are lacking. Methods: Consecutive patients referred for surveillance endoscopy for Barrett’s esophagus underwent high-definition white light and narrow band imaging followed by pCLE and targeted biopsy or mucosal resection. Subsequently, Seattle protocol biopsies were obtained during the same session. Baseline variables, real-time pCLE interpretation, and histology results were prospectively recorded. An expert endomicroscopist performed blinded review of pCLE sequences, and an expert pathologist performed blinded review of histologic specimens. A sample size of 64 patients was calculated a priori based on 3% estimated prevalence of high-grade dysplasia (HGD) or cancer, 80% per-patient pCLE specificity for HGD or cancer, power 80%, and alpha 0.05. Results: 66 patients were included in the study (mean age 63 years, 98% male, and 92% white). Mean Barrett’s segment length was C2M3. The median time to perform pCLE was 7 minutes. The prevalence of HGD or cancer was 4.47% (2 cancers, 1 HGD). Both real-time and blinded pCLE correctly identified all cases of cancer. For the primary outcome, accuracy of real-time pCLE for diagnosing HGD/cancer compared to non-blinded pathologist interpretation was as follows: sensitivity 66.7%, specificity 96.8%, negative predictive value 50%, and positive predictive value 98.4% (Table 1). For HGD and cancer, inter-observer agreement was substantial between realtime and blinded endomicroscopists (kappaZ0.6), and was perfect between blinded and non-blinded pathologists (kappaZ1). For low-grade dysplasia interobserver agreement was minimal between endomicroscopists as well as between pathologists (kappaZ0.2 in both groups; Table 2). Conclusions: pCLE demonstrates high specificity for detecting dysplasia, but lower sensitivity may limit its widespread adoption as a replacement to random biopsies. There is substantial disagreement between endomicroscopists and between pathologists for the
www.giejournal.org
Abstracts
diagnosis of low-grade dysplasia. Cost-effectiveness analyses may help identify the optimal role for this technology in patients with Barrett’s esophagus.
Table 1: Accuracy of probe-based confocal endomicroscopy (pCLE) compared to histology for high-grade dysplasia or cancer
Real-time pCLE interpretation
Blinded pCLE interpretation
Non-blinded pathologist interpretation
Blinded pathologist interpretation
Sensitivity 66.7% Specificity 96.8% NPV 50% PPV 98.4% Sensitivity 66.7% Specificity 93.7% NPV 33.3% PPV 98.3%
Sensitivity 66.7% Specificity 96.8% NPV 50% PPV 98.4% Sensitivity 66.7% Specificity 93.7% NPV 33.3% PPV 98.3%
PPV – positive predictive value; NPV – negative predictive value.
Table 2: Accuracy of probe-based confocal endomicroscopy (pCLE) compared to histology for low-grade dysplasia
Real-time pCLE interpretation
Blinded pCLE interpretation
Non-blinded pathologist interpretation
Blinded pathologist interpretation
Sensitivity 60% Specificity 86.9% NPV 27.3% PPV 96.4% Sensitivity 0% Specificity 93.4% NPV 0% PPV 91.9%
Sensitivity 31.6% Specificity 89.4% NPV 54.5% PPV 76.4% Sensitivity 10.5% Specificity 95.7% NPV 50% PPV 72.6%
PPV – positive predictive value; NPV – negative predictive value.
537 Quality of Endoscopic Surveillance of Barrett’s Esophagus in Denmark Jes S. Vogt*1, Anders Christian Larsen1, Thorbjørn Sommer2, Per Ejstrud1 1 Department of Gastrointestinal Surgery, Aalborg University Hospital, Aalborg, Denmark; 2Department of Surgery, Region Hospital Randers, Randers, Denmark Background: The increasing incidence of esophageal adenocarcinoma and novel endoluminal treatments of Barrett’s neoplasia, emphasizes the importance of high quality surveillance of Barrett’s esophagus (BE). Adherence to surveillance guidelines in Denmark has never been investigated. The aim of this study was to perform a retrospective national study evaluating the compliance with Danish national guidelines. Methods: During a three year period, a total of 3692 patients with BE were identified using The Danish Pathology Registry. 300 patients were included by drawing a simple random sample. Information regarding description of the BE segment, biopsy protocol, communication with the pathologist and planned followup endoscopy, was obtained. Results: From the subset of 300 patients, we excluded 31 patients (10%) due to missing reports. 83 patients (28%) had no endoscopic evidence of BE according to the endoscopy reports and were also excluded. BE was macroscopically suspected by the endoscopist in 186 patients (62%) and these patients were included in this study. Prague C&M classification was recorded in 63 of the endoscopy reports (34%). The BE segment was stratified by lengths of 1-5 cm, 510 cm and 10-15 cm and the median number of biopsy specimens obtained in each segment was 4 (interquartile range (IQR), 3-6), 5.5 (IQR, 4-9) and 4 (IQR, 1-6). According to the modified Seattle biopsy protocol in the Danish guidelines, a correct minimum number of biopsies were obtained in only 12%, 8% and 0% of cases, respectively. The median number of biopsies in all patients with endoscopic evidence of BE was 4 (IQR 3-6). 95% of endoscopists reported in the pathology requisition that the indication for histological evaluation was BE. 28% described the exact location of the biopsies in the pathology requisition. Patients with BE and no dysplasia, had endoscopic surveillance performed after a median of 24 months (IQR, 6-24). Conclusion: Adherence to Danish guidelines is poor and may be associated with insufficient quality of BE surveillance and reduced detection of neoplasia in BE patients. The substantial lack of endoscopic evidence of BE and non-compliance to guidelines in our study, call for caution when interpreting and conducting studies based on pathology registry. Previously published studies based on The Danish Pathology Registry and the quality of endoscopic surveillance may substantially underestimate the incidence of adenocarcinoma in BE patients.
www.giejournal.org
538 Feasibility and Performance Characteristics of a Novel Disposable Transnasal Capsule Device for Barrett’s Esophagus Screening: A Prospective International Multicenter Trial Prachi Pophali1, Sami Sarmed2, Magnus Halland3, Massimiliano DiPietro3, Felicity Enders1, Jacobo Ortiz Fernández-Sordo2, Jonathan White2, Michele L. Johnson1, Indra Guha2, Rebecca C. Fitzgerald3, Krish Ragunath2, Prasad G. Iyer*1 1 Mayo Clinic, Rochester, MN; 2Queens Medical Centre, Nottingham, United Kingdom; 3University of Cambridge, Cambridge, United Kingdom Background: Screening for Barrett’s esophagus (BE) with standard endoscopy (SE) is expensive. Unsedated transnasal endoscopy (uTNE) is a feasible and safe alternative for BE screening. We aimed to assess the technical feasibility, safety and performance characteristics of clinic-based uTNE using a novel single use transnasal capsule based device (EG Scan II) compared to SE. Methods: A prospective study was conducted at three centers (2 United Kingdom, 1 United States). Patients referred for clinical SE with and without known BE were invited to participate. uTNE with the EG Scan device (consisting of a 6 mm diameter disposable capsule attached to a controller and air compressor) was performed first followed by SE. Both procedures were conducted by different endoscopists blinded to the indication and outcome of either procedure. Results of the SE were considered as the gold standard. Patient symptoms (gagging, choking, discomfort, nasal pain) were recorded on a 10-point visual analogue scale (VAS) (10 being the worst), in addition to overall preference and tolerability on a 10-point VAS (10 being the best). Results: 200 patients, 100 with and 100 without known BE, 33% females and mean (SD) age of 57.8 (14) years participated. Median (IQR) length of the BE segment was 2 (1, 4) cm. 178 (89%) patients completed both procedures. 22 (11%) did not complete the EG Scan exam (1 poor tolerability after intubation, 21 inability to traverse nasal passages). The median symptom scores for the EG Scan were excellent ranging from 0-2. Overall tolerability was good with a median VAS score of 8 for the EG Scan, compared to 7 for SE. 54.5% of the participants preferred the EG Scan, while 16.8% preferred SE. 1 patient developed self-limiting epistaxis following uTNE. Sensitivity and specificity of EG Scan for BE diagnosis was 89.4% (95% CI 83.3- 95.6) and 90.3% (95% CI 84 – 96.7) respectively. Sensitivity was superior for long segment BE (93%) than short segment BE (87%). Conclusions: uTNE with the EG Scan device is a safe, welltolerated and accurate alternative for BE screening. Thus, it could potentially serve as a community and office based tool for BE screening.
539 Dysplasia Recurrence After Complete Eradication of Intestinal Metaplasia (CEIM) and Dysplasia (CED) in Barrett’s Esophagus Treated With Endoscopic Therapy: Results From an International, Multi-Center Consortium Rajesh Krishnamoorthi*1, Sreekar Vennelaganti9, Alessandro Repici8, Vani J. Konda3, Irving Waxman3, Stefan Seewald4, Matthew W. Stier3, Rehan Haidry6, Daniel Buckles2, Neil Gupta5, Ajay Bansal2, Sharad Mathur2, Mojtaba S. Olyaee2, Michael J. Bourke7, Gary W. Falk10, Ramprasad Jegadeesan2, Julie Nguyen2, Prashanth Vennalaganti9, Anusha Vittal2, Kevin F. Kennedy9, Prateek Sharma2,9, Andrew S. Ross1 1 Digestive Disease Institute, Virginia Mason Medical Center, Seattle, WA; 2 University of Kansas, Kansas City, KS; 3University of Chicago, Chicago, IL; 4Hirsladen, Zurich, Switzerland; 5Loyola University, Maywood, IL; 6 University college London Hospital, London, United Kingdom; 7 University of Sydney, Westmead, New South Wales, Australia; 8Instituto Clinico Humanitas, Milan, Italy; 9Kansas city VA Medical Center, Kansas city, MO; 10University of Pennsylvania Perelman School of Medicine, Philadelphia, PA Background: CEIM is more difficult to achieve than CED during endoscopic therapy for dysplastic BE. Hence, it is important to define the goal of endoscopic therapy for each individual patient – complete elimination of intestinal metaplasia (CEIM) or dysplasia (CED). We aimed to compare the risk of recurrent dysplastic BE in patients who achieved CEIM and CED. Methods: We reviewed all BE subjects who underwent endoscopic therapy in a multicenter prospective registry. BE subjects with LGD, HGD and EAC who reached CEIM or CED were included in this analysis. Dysplastic recurrence in the CEIM and CED groups was defined as presence of recurrent BE endoscopically (columnar mucosa in distal esophagus) with histologic evidence of any dysplasia or EAC after achieving CEIM and CED respectively. Cox proportional hazard models were used to compare the risk of dysplastic recurrence between CEIM and CED groups. Results: 266 subjects with dysplastic BE who underwent endoscopic therapy were included in the study. 191 subjects achieved CEIM and 75 subjects achieved CED. The mean (SD) age was 64 (11.2) years.240 (90.2%) were males. The median follow up of subjects was 31.2 months. Overall, 52 subjects (19.5%) had recurrence of dysplasia; the rate of recurrence was 21% in the CEIM group and 16% in the in the CED group. Baseline characteristics of CEIM and CED groups are summarized in table 1. The risk of dysplastic recurrence in CEIM group
Volume 85, No. 5S : 2017 GASTROINTESTINAL ENDOSCOPY AB77