- Email: [email protected]
T1879 Low-Grade Dysplasia (LGD) Is a Poor Marker for Cancer Progression in Patients with Barrett's Esophagus (BE): Preliminary Results from a Large, Multi-Center, Cohort Study
AGA Abstracts
(OR 2.72, CI 1.46-5.06) and esophageal body hypomotility (OR 1.90, CI 1.18-3.07) were independent predictors of BE. CONCLUSIONS: Using HRM, BE is intimately associated with esophageal body and LES hypomotility; both are independent predictors of BE in the setting of reflux disease. The association with LSBE is more robust than that with SSBE.
T1881 Detection of Dysplasia in Barrett's Esophagus with Angle-Resolved Low Coherence Interferometry Adam Wax, Neil G. Terry, Yizheng Zhu, Steven C. Gebhart, William J. Brown, Stephanie D. Bright, Elizabeth E. Carretta, John T. Woosley, Nicholas J. Shaheen BACKGROUND: Traditional endoscopic surveillance of Barrett's esophagus (BE) suffers from a high rate of sampling error. Angle-resolved low coherence interferometry (a/LCI) is a noninvasive method of assessing neoplastic tissue. a/LCI obtains nuclear morphology measurements of epithelial tissues by combining the depth resolution of optical coherence tomography with the sensitivity to nuclear morphology of light scattering spectroscopy. We report the first In Vivo application of a/CLI to assess BE. METHODS: BE patients with low-grade dysplasia (LGD) or non-dysplastic BE undergoing standard surveillance were enrolled to compare dysplasia yield between a/LCI and standard biopsy. In addition to routine surveillance biopsies, additional points on the epithelium were randomly selected for imaging with the a/LCI probe. A biopsy was taken from each imaged point, guided by the transient tissue indentation at the probe site. . Data were acquired by actuating a foot pedal when the probe was in apposition to a region of interest. The primary endpoints were the sensitivity, specificity and overall accuracy of the a/LCI probe for identifying dysplasia. We will generate kappa values for agreement between the biopsies and images upon completion of the data set. RESULTS: Fifty-two paired biopsies and a/LCI images were taken in 23 enrolled patients to date. Overall the a/LCI probe achieved 100% sensitivity (8/8 samples) and 84% (26/31) specificity for distinguishing dysplasia from non-dysplastic tissue. In addition, the a/LCI probe identified 11 points as dysplastic which were classified as “indeterminate for dysplasia” by the study pathologist. Significantly, even though the paired study biopsy was read as indeterminant in these instances, the non-study biopsies revealed LGD in all subjects. No points identified as indeterminate were labeled as non-dysplastic by the a/LCI probe. CONCLUSION: In this first In Vivo pilot study, a/LCI demonstrated good sensitivity and specificity for detecting dysplasia in BE. As additional data become available, the performance characteristics of a/LCI will be better elucidated.
T1879 Low-Grade Dysplasia (LGD) Is a Poor Marker for Cancer Progression in Patients with Barrett's Esophagus (BE): Preliminary Results from a Large, Multi-Center, Cohort Study Sachin B. Wani, Gary W. Falk, Richard E. Sampliner, Amy Wang, Patrick E. Young, David A. Lieberman, Ajay Bansal, Brooks D. Cash, Mandeep Singh, Vikas Singh, April D. Higbee, Matthew Hall, Amit Rastogi, Prateek Sharma Background: Data on neoplastic progression of LGD in BE are highly variable [cancer incidence rates (IR): 0.6-2.6%/year]. Aim:In a large multi-center cohort: - determine IR of high-grade dysplasia (HGD)/EAC in BE pts with LGD - determine IR of HGD/EAC in pts with prevalent vs incident LGD; and in those with LGD on more than 1 occasion. Methods:This is a prospective, multicenter outcomes project involving BE pts from 5 centers. Demographic information, endoscopic information and histologic diagnosis were recorded. The duration of f/u was calculated from the time of initial BE diagnosis to the most recent EGD. The time of occurrence of LGD, HGD, and EAC was documented which allowed determination of HGD/EAC incidence. Data from each center were merged. BE pts with LGD at initial EGD (prevalent LGD) and f/u of at least 12 mths and BE pts with no dysplasia at initial EGD who developed LGD after a follow-up of at least 12 mths (incident LGD)were identified. Persistent LGD was defined as LGD documented at least twice, either on consecutive or non-consecutive EGDs. Incidence of HGD and EAC was calculated in these cohorts. Results:In this analysis including pts from 3 centers, 245 BE pts with LGD f/u of at least 12 mths were identified mean age 62.4 yrs, 98.2% Caucasians and 85.6% men). The mean BE length was 4.3 cm (range: 1-7) and mean f/u was 4.5 years (range: 2-5.9). 27 pts developed HGD and 7 developed EAC. The annual IR of HGD was 2.4% (95% CI 1.7-3.6) and that for EAC was 0.6% (95% CI 0.2-1.2). 113 cases of prevalent LGD were identified with a mean f/u of 4.5 yrs (range: 1.9-6). 11 of these developed HGD and 2 developed EAC - annual IR of HGD was 2.1% (95% CI 1.8-3.8), and that for EAC was 0.4% (95% CI 0.94-1.5). 97 patients with incident LGD were identified with a mean follow-up of 4.6 years (range: 2-6). In this group, 10 patients developed HGD and 4 developed - the annual IR of HGD was 2.2% (95% CI 1.2-4.2), and for EAC was 0.8% (95% CI 0.3-2.2). The IR of EAC in pts with persistent LGD documented on 2 consecutive occasions [annual IR: 0.9% (95% CI 0.3-2.9)] and in those with LGD documented twice at any point in their follow up [annual IR: 0.9% (95% CI 0.3-2.5)] was not significantly higher compared to the overall LGD group. Conclusions:Progression to EAC for LGD pts in this cohort was 0.6%/year; similar to published rates of progression in non-dysplastic BE(0.5%/yr). There was no significant difference in the incidence of HGD and EAC in pts with prevalent vs incident LGD; or in those with persistent LGD. LGD appears to be a poor marker for EAC development; further research to help identify a high risk group for development of HGD/EAC is needed.
T1882 Buried Barrett After Radiofrequency Ablation for Neoplastic Barrett Esophagus: Undetectable Due to Mucosal Scarring or Truly a Rare Occurrence? Roos E. Pouw, Carine Sondermeijer, Fiebo J. ten Kate, Robert D. Odze, Michael Vieth, Jacques Bergman BACKGROUND: Radiofrequency ablation (RFA) is safe and effective for eradicating neoplastic Barrett esophagus (BE). Although buried Barrett (BB) glands underneath the neosquamous epithelium (NSE) are an extremely rare finding during follow-up, some have hypothesized that BB cannot be adequately sampled due to presumptive mucosal fibrosis after RFA. AIMS: Prospectively evaluate sampling depth of primary and keyhole biopsies obtained from NSE vs. untreated squamous epithelium (USE) as control. Compare sampling depth of standard vs. jumbo biopsy forceps in NSE and USE. Assess for BB beneath the NSE using primary biopsy, keyhole biopsy and endoscopic resection (ER). METHODS: We considered 23 patients for enrollment, all of whom had undergone RFA for neoplastic BE under protocol. After signing informed consent, patients were randomized to undergo standard vs. jumbo biopsies from the NSE and USE (4Q/2cm). After each primary biopsy a “keyhole” biopsy was obtained from the same site in order to obtain deeper tissue. In addition an ER specimen was obtained from an area of NSE. Three expert pathologists independently scored (blinded to biopsy source) histological depth for each biopsy and ER specimen and determined if BB was present. RESULTS: 16 of 23 patients participated (exclusions: unrelated death (1), co-morbidity (2), initial BE <2 cm (4)). Complete eradication of neoplastic BE had been achieved and sustained in all patients prior to enrollment; median follow-up 26 months (IQR 21-28). There was no difference in primary biopsy depth between NSE vs. USE: lamina propria was sampled in 37% and 36% of cases, respectively. Keyhole biopsies sampled significantly deeper than primary non-keyhole biopsies. There was no significant difference in sampling depth between standard and jumbo biopsy forceps. All ER-specimens included submucosa. BB was not found in any of the (keyhole) biopsies and ER specimens. CONCLUSION: Primary biopsies from NSE after RFA sample as deeply as biopsies from control USE. It is thus unlikely that the reported absence of BB after RFA reflects insufficient biopsy depth due to mucosal scarring. We found no benefit in terms of additional biopsy depth for using jumbo biopsy forceps in surveying post-RFA patients. The absence of BB in (keyhole) biopsies and ER specimens suggests that RFA obliterates all Barrett mucosa, both superficial and deep.
T1880 Esophageal Multilayered Epithelium Is a Dynamic Epithelium with Bidirectional (Squamous and Columnar) Differentiation Capability Mari Mino-Kenudson, Amitabh Srivastava, Jonathan N. Glickman, Robert D. Odze Background: Multilayered epithelium (ME) is a hybrid epithelium believed to represent a transitional step in the conversion of squamous to columnar epithelium in Barrett's esophagus (BE). However, some evidence suggests that ME may also represent an intermediate phase in columnar to squamous metaplasia in certain circumstances, such as in the restitution of squamous epithelium after photodynamic therapy (PDT). The aim of this study was to evaluate and compare the biological properties of ME in 3 different biological settings: 1) BE patients on proton pump inhibitor (PPI) therapy; 2) BE patients post-PDT; and 3) fetal esophagus in which columnar epithelium converts to squamous epithelium during development in-utero. Design: Routinely processed biopsies with ME from 14 BE patients on PPI therapy, 10 patients (12 biopsies) with BE post-PDT, and 16 fetuses and neonates (gestational age: between 21- 41 weeks) were evaluated by immunohistochemistry for a panel of markers of intestinal, gastric and squamous differentiation (MUC2, MUC5AC, MUC5B, MUC6, Cdx-2 and p63), and proliferation (Ki-67). The findings in ME were compared between the 3 study groups. Result: Both p63 (a squamous epithelium biomarker) and MUC5B (a marker for esophageal gland/ducts) showed a significantly increased prevalence rate and intensity of staining in post-PDT cases (100% and 100%, respectively) compared to BE patients post PPI (50% and 43%, respectively). Expression of both p63 and MUC5B were prevalent in fetuses (81% and 100%, respectively), but the intensity of MUC5B staining was significantly less than in post-PDT cases (p = 0.03). In contrast, the intestinal columnar epithelium biomarkers Cdx-2 and MUC2 were expressed in significantly fewer fetal cases (13% and 19%, respectively) compared to BE patients on PPI therapy (50% and 43%, respectively) and BE patients post-PDT (67% and 67%, respectively). No differences were observed with regard to MUC5AC, MUC6 staining or the Ki-67 proliferation rate in ME between the 3 groups. Conclusion: ME is a dynamic epithelium that probably represents an intermediate step in not only the squamous-columnar transition in BE, but also in the process of restoration of native squamous mucosa in patients post PDT who are reverting from a columnar to a squamous phenotype. Our findings suggest the existence of, as yet unidentified, cells with pluripotent capability of bidirectional differentiation, possibly located in the esophageal gland/ducts, in response to external stimuli.
AGA Abstracts
A-592
(OR 2.72, CI 1.46-5.06) and esophageal body hypomotility (OR 1.90, CI 1.18-3.07) were independent predictors of BE. CONCLUSIONS: Using HRM, BE is intimately associated with esophageal body and LES hypomotility; both are independent predictors of BE in the setting of reflux disease. The association with LSBE is more robust than that with SSBE.
T1881 Detection of Dysplasia in Barrett's Esophagus with Angle-Resolved Low Coherence Interferometry Adam Wax, Neil G. Terry, Yizheng Zhu, Steven C. Gebhart, William J. Brown, Stephanie D. Bright, Elizabeth E. Carretta, John T. Woosley, Nicholas J. Shaheen BACKGROUND: Traditional endoscopic surveillance of Barrett's esophagus (BE) suffers from a high rate of sampling error. Angle-resolved low coherence interferometry (a/LCI) is a noninvasive method of assessing neoplastic tissue. a/LCI obtains nuclear morphology measurements of epithelial tissues by combining the depth resolution of optical coherence tomography with the sensitivity to nuclear morphology of light scattering spectroscopy. We report the first In Vivo application of a/CLI to assess BE. METHODS: BE patients with low-grade dysplasia (LGD) or non-dysplastic BE undergoing standard surveillance were enrolled to compare dysplasia yield between a/LCI and standard biopsy. In addition to routine surveillance biopsies, additional points on the epithelium were randomly selected for imaging with the a/LCI probe. A biopsy was taken from each imaged point, guided by the transient tissue indentation at the probe site. . Data were acquired by actuating a foot pedal when the probe was in apposition to a region of interest. The primary endpoints were the sensitivity, specificity and overall accuracy of the a/LCI probe for identifying dysplasia. We will generate kappa values for agreement between the biopsies and images upon completion of the data set. RESULTS: Fifty-two paired biopsies and a/LCI images were taken in 23 enrolled patients to date. Overall the a/LCI probe achieved 100% sensitivity (8/8 samples) and 84% (26/31) specificity for distinguishing dysplasia from non-dysplastic tissue. In addition, the a/LCI probe identified 11 points as dysplastic which were classified as “indeterminate for dysplasia” by the study pathologist. Significantly, even though the paired study biopsy was read as indeterminant in these instances, the non-study biopsies revealed LGD in all subjects. No points identified as indeterminate were labeled as non-dysplastic by the a/LCI probe. CONCLUSION: In this first In Vivo pilot study, a/LCI demonstrated good sensitivity and specificity for detecting dysplasia in BE. As additional data become available, the performance characteristics of a/LCI will be better elucidated.
T1879 Low-Grade Dysplasia (LGD) Is a Poor Marker for Cancer Progression in Patients with Barrett's Esophagus (BE): Preliminary Results from a Large, Multi-Center, Cohort Study Sachin B. Wani, Gary W. Falk, Richard E. Sampliner, Amy Wang, Patrick E. Young, David A. Lieberman, Ajay Bansal, Brooks D. Cash, Mandeep Singh, Vikas Singh, April D. Higbee, Matthew Hall, Amit Rastogi, Prateek Sharma Background: Data on neoplastic progression of LGD in BE are highly variable [cancer incidence rates (IR): 0.6-2.6%/year]. Aim:In a large multi-center cohort: - determine IR of high-grade dysplasia (HGD)/EAC in BE pts with LGD - determine IR of HGD/EAC in pts with prevalent vs incident LGD; and in those with LGD on more than 1 occasion. Methods:This is a prospective, multicenter outcomes project involving BE pts from 5 centers. Demographic information, endoscopic information and histologic diagnosis were recorded. The duration of f/u was calculated from the time of initial BE diagnosis to the most recent EGD. The time of occurrence of LGD, HGD, and EAC was documented which allowed determination of HGD/EAC incidence. Data from each center were merged. BE pts with LGD at initial EGD (prevalent LGD) and f/u of at least 12 mths and BE pts with no dysplasia at initial EGD who developed LGD after a follow-up of at least 12 mths (incident LGD)were identified. Persistent LGD was defined as LGD documented at least twice, either on consecutive or non-consecutive EGDs. Incidence of HGD and EAC was calculated in these cohorts. Results:In this analysis including pts from 3 centers, 245 BE pts with LGD f/u of at least 12 mths were identified mean age 62.4 yrs, 98.2% Caucasians and 85.6% men). The mean BE length was 4.3 cm (range: 1-7) and mean f/u was 4.5 years (range: 2-5.9). 27 pts developed HGD and 7 developed EAC. The annual IR of HGD was 2.4% (95% CI 1.7-3.6) and that for EAC was 0.6% (95% CI 0.2-1.2). 113 cases of prevalent LGD were identified with a mean f/u of 4.5 yrs (range: 1.9-6). 11 of these developed HGD and 2 developed EAC - annual IR of HGD was 2.1% (95% CI 1.8-3.8), and that for EAC was 0.4% (95% CI 0.94-1.5). 97 patients with incident LGD were identified with a mean follow-up of 4.6 years (range: 2-6). In this group, 10 patients developed HGD and 4 developed - the annual IR of HGD was 2.2% (95% CI 1.2-4.2), and for EAC was 0.8% (95% CI 0.3-2.2). The IR of EAC in pts with persistent LGD documented on 2 consecutive occasions [annual IR: 0.9% (95% CI 0.3-2.9)] and in those with LGD documented twice at any point in their follow up [annual IR: 0.9% (95% CI 0.3-2.5)] was not significantly higher compared to the overall LGD group. Conclusions:Progression to EAC for LGD pts in this cohort was 0.6%/year; similar to published rates of progression in non-dysplastic BE(0.5%/yr). There was no significant difference in the incidence of HGD and EAC in pts with prevalent vs incident LGD; or in those with persistent LGD. LGD appears to be a poor marker for EAC development; further research to help identify a high risk group for development of HGD/EAC is needed.
T1882 Buried Barrett After Radiofrequency Ablation for Neoplastic Barrett Esophagus: Undetectable Due to Mucosal Scarring or Truly a Rare Occurrence? Roos E. Pouw, Carine Sondermeijer, Fiebo J. ten Kate, Robert D. Odze, Michael Vieth, Jacques Bergman BACKGROUND: Radiofrequency ablation (RFA) is safe and effective for eradicating neoplastic Barrett esophagus (BE). Although buried Barrett (BB) glands underneath the neosquamous epithelium (NSE) are an extremely rare finding during follow-up, some have hypothesized that BB cannot be adequately sampled due to presumptive mucosal fibrosis after RFA. AIMS: Prospectively evaluate sampling depth of primary and keyhole biopsies obtained from NSE vs. untreated squamous epithelium (USE) as control. Compare sampling depth of standard vs. jumbo biopsy forceps in NSE and USE. Assess for BB beneath the NSE using primary biopsy, keyhole biopsy and endoscopic resection (ER). METHODS: We considered 23 patients for enrollment, all of whom had undergone RFA for neoplastic BE under protocol. After signing informed consent, patients were randomized to undergo standard vs. jumbo biopsies from the NSE and USE (4Q/2cm). After each primary biopsy a “keyhole” biopsy was obtained from the same site in order to obtain deeper tissue. In addition an ER specimen was obtained from an area of NSE. Three expert pathologists independently scored (blinded to biopsy source) histological depth for each biopsy and ER specimen and determined if BB was present. RESULTS: 16 of 23 patients participated (exclusions: unrelated death (1), co-morbidity (2), initial BE <2 cm (4)). Complete eradication of neoplastic BE had been achieved and sustained in all patients prior to enrollment; median follow-up 26 months (IQR 21-28). There was no difference in primary biopsy depth between NSE vs. USE: lamina propria was sampled in 37% and 36% of cases, respectively. Keyhole biopsies sampled significantly deeper than primary non-keyhole biopsies. There was no significant difference in sampling depth between standard and jumbo biopsy forceps. All ER-specimens included submucosa. BB was not found in any of the (keyhole) biopsies and ER specimens. CONCLUSION: Primary biopsies from NSE after RFA sample as deeply as biopsies from control USE. It is thus unlikely that the reported absence of BB after RFA reflects insufficient biopsy depth due to mucosal scarring. We found no benefit in terms of additional biopsy depth for using jumbo biopsy forceps in surveying post-RFA patients. The absence of BB in (keyhole) biopsies and ER specimens suggests that RFA obliterates all Barrett mucosa, both superficial and deep.
T1880 Esophageal Multilayered Epithelium Is a Dynamic Epithelium with Bidirectional (Squamous and Columnar) Differentiation Capability Mari Mino-Kenudson, Amitabh Srivastava, Jonathan N. Glickman, Robert D. Odze Background: Multilayered epithelium (ME) is a hybrid epithelium believed to represent a transitional step in the conversion of squamous to columnar epithelium in Barrett's esophagus (BE). However, some evidence suggests that ME may also represent an intermediate phase in columnar to squamous metaplasia in certain circumstances, such as in the restitution of squamous epithelium after photodynamic therapy (PDT). The aim of this study was to evaluate and compare the biological properties of ME in 3 different biological settings: 1) BE patients on proton pump inhibitor (PPI) therapy; 2) BE patients post-PDT; and 3) fetal esophagus in which columnar epithelium converts to squamous epithelium during development in-utero. Design: Routinely processed biopsies with ME from 14 BE patients on PPI therapy, 10 patients (12 biopsies) with BE post-PDT, and 16 fetuses and neonates (gestational age: between 21- 41 weeks) were evaluated by immunohistochemistry for a panel of markers of intestinal, gastric and squamous differentiation (MUC2, MUC5AC, MUC5B, MUC6, Cdx-2 and p63), and proliferation (Ki-67). The findings in ME were compared between the 3 study groups. Result: Both p63 (a squamous epithelium biomarker) and MUC5B (a marker for esophageal gland/ducts) showed a significantly increased prevalence rate and intensity of staining in post-PDT cases (100% and 100%, respectively) compared to BE patients post PPI (50% and 43%, respectively). Expression of both p63 and MUC5B were prevalent in fetuses (81% and 100%, respectively), but the intensity of MUC5B staining was significantly less than in post-PDT cases (p = 0.03). In contrast, the intestinal columnar epithelium biomarkers Cdx-2 and MUC2 were expressed in significantly fewer fetal cases (13% and 19%, respectively) compared to BE patients on PPI therapy (50% and 43%, respectively) and BE patients post-PDT (67% and 67%, respectively). No differences were observed with regard to MUC5AC, MUC6 staining or the Ki-67 proliferation rate in ME between the 3 groups. Conclusion: ME is a dynamic epithelium that probably represents an intermediate step in not only the squamous-columnar transition in BE, but also in the process of restoration of native squamous mucosa in patients post PDT who are reverting from a columnar to a squamous phenotype. Our findings suggest the existence of, as yet unidentified, cells with pluripotent capability of bidirectional differentiation, possibly located in the esophageal gland/ducts, in response to external stimuli.
AGA Abstracts
A-592