Clinical utility and interobserver agreement of autofluorescence imaging and magnification narrow-band imaging for the evaluation of Barrett's esophagus: a prospective tandem study

Background New endoscopic imaging techniques, such as autofluorescence imaging (AFI) and narrow-band imaging (NBI), have been developed to improve the detection of neoplastic lesions in Barrett's esophagus (BE). Objective To evaluate the clinical utility of AFI and magnification NBI to detect high-grade dysplasia (HGD) and early esophageal adenocarcinoma (EAC) and the interobserver agreement. Design Prospective tandem study of eligible patients. Setting Single, academic tertiary care center. Patients Forty-two patients with a history of confirmed BE were prospectively enrolled. Interventions The BE segment was examined under high-definition white-light endoscopy, and the presence of visible lesions was recorded. Subsequently, AFI and magnification NBI were performed in tandem on areas of the BE segment away from visible lesions; images obtained by these 2 systems were graded according to the color of reflected light and surface patterns, respectively. Biopsy specimens were obtained at the end of the procedure. Main Outcome Measurements The sensitivity, specificity, positive predictive value, and negative predictive value (NPV) of the AFI and NBI patterns for the detection of HGD/EAC and interobserver agreement. Results Of the 42 patients enrolled, 14 (33%) had HGD/EAC. On patient-based analysis, AFI alone had a sensitivity, specificity, and NPV of 50%, 61%, and 71%, respectively, and the overall accuracy for the detection of HGD/EAC patients was 57%. By using magnification NBI in tandem fashion, the sensitivity and NPV improved to 71% and 76%, respectively, with a decrease in specificity to 46% and in overall accuracy to 55%. The 2 techniques had moderate interobserver agreement for both the patterns and prediction of histology. Limitations Uncontrolled study performed at an academic center by expert endoscopists in a high-risk population. Conclusions By using a multimodality endoscope, both AFI and magnification NBI had limited clinical accuracy and moderate overall interobserver agreement. AFI does not appear to be useful as a broad-based technique for the detection of neoplasia in patients with BE.