TY - JOUR
T1 - Identification of predictive factors for early neoplasia in Barrett's esophagus after autofluorescence imaging
T2 - a stepwise multicenter structured assessment
AU - Curvers, Wouter L.
AU - Singh, Rajvinder
AU - Wallace, Michael B.
AU - Wong Kee Song, Louis Michel
AU - Ragunath, Krish
AU - Wolfsen, Herbert C.
AU - ten Kate, Fiebo J.
AU - Fockens, Paul
AU - Bergman, Jacques J.G.H.M.
N1 - Funding Information:
DISCLOSURE: The following authors disclosed financial relationships relevant to this publication: W. Curvers is financially supported by an unrestricted research grant from Astra Zeneca, the Netherlands. P. Fockens and J. Bergman have received research grants from Olympus Inc, Tokyo, Japan. All other authors disclosed no financial relationships relevant to this publication.The department of Gastroenterology and Hepatology of the Academic Medical Centre Amsterdam has been provided with loaned endoscopic equipment by Olympus Inc, Tokyo, Japan.
Copyright:
Copyright 2009 Elsevier B.V., All rights reserved.
PY - 2009/7
Y1 - 2009/7
N2 - Background: Autofluorescence imaging is a novel imaging technique that may improve the detection of early neoplasia in Barrett's esophagus. Autofluorescence imaging is, however, associated with a 40% to 81% false-positive rate. Objective: Our purpose was to identify endoscopic features that may predict the presence of early neoplasia in autofluorescence-positive areas. Design: Descriptive and prospective cohort study. Setting: Tertiary referral centers for the detection and treatment of early Barrett's neoplasia. Patients and Methods: Patients undergoing autofluorescence endoscopy. High-quality images with autofluorescence imaging and white-light endoscopy were obtained with corresponding histologic study. A systematic image evaluation process was performed, including an unblinded orientation phase (10 areas), a blinded derivation phase, and a blinded validation phase by 5 international experts in autofluorescence imaging (80 areas). Subsequently the identified features were validated in a prospective pilot study. Main Outcome Measurements: Association between endoscopic features and presence of early neoplasia in autofluorescence-positive areas. Results: Autofluorescence intensity, proximity of gastric folds <1 cm, and different appearance on white-light endoscopy were independently associated with early neoplasia in autofluorescence-positive areas on multivariate analysis. The κ values for interobserver agreement of these factors were moderate, ranging between 0.49 to 0.56. The association with autofluorescence intensity and different appearance on white-light endoscopy was confirmed in a prospective pilot study. Limitation: Selected set of images from a high-risk population (tertiary referral center). Conclusion: We found specific endoscopic features that were associated with early neoplasia in autofluorescence-positive areas. These findings can be used in future prospective studies to improve the accuracy of autofluorescence imaging without performing magnification endoscopy for detailed inspection of suspicious areas.
AB - Background: Autofluorescence imaging is a novel imaging technique that may improve the detection of early neoplasia in Barrett's esophagus. Autofluorescence imaging is, however, associated with a 40% to 81% false-positive rate. Objective: Our purpose was to identify endoscopic features that may predict the presence of early neoplasia in autofluorescence-positive areas. Design: Descriptive and prospective cohort study. Setting: Tertiary referral centers for the detection and treatment of early Barrett's neoplasia. Patients and Methods: Patients undergoing autofluorescence endoscopy. High-quality images with autofluorescence imaging and white-light endoscopy were obtained with corresponding histologic study. A systematic image evaluation process was performed, including an unblinded orientation phase (10 areas), a blinded derivation phase, and a blinded validation phase by 5 international experts in autofluorescence imaging (80 areas). Subsequently the identified features were validated in a prospective pilot study. Main Outcome Measurements: Association between endoscopic features and presence of early neoplasia in autofluorescence-positive areas. Results: Autofluorescence intensity, proximity of gastric folds <1 cm, and different appearance on white-light endoscopy were independently associated with early neoplasia in autofluorescence-positive areas on multivariate analysis. The κ values for interobserver agreement of these factors were moderate, ranging between 0.49 to 0.56. The association with autofluorescence intensity and different appearance on white-light endoscopy was confirmed in a prospective pilot study. Limitation: Selected set of images from a high-risk population (tertiary referral center). Conclusion: We found specific endoscopic features that were associated with early neoplasia in autofluorescence-positive areas. These findings can be used in future prospective studies to improve the accuracy of autofluorescence imaging without performing magnification endoscopy for detailed inspection of suspicious areas.
UR - http://www.scopus.com/inward/record.url?scp=67649250017&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=67649250017&partnerID=8YFLogxK
U2 - 10.1016/j.gie.2008.10.026
DO - 10.1016/j.gie.2008.10.026
M3 - Article
C2 - 19394009
AN - SCOPUS:67649250017
SN - 0016-5107
VL - 70
SP - 9
EP - 17
JO - Gastrointestinal endoscopy
JF - Gastrointestinal endoscopy
IS - 1
ER -