Inter- and intra-observer variability of magnification chromoendoscopy for detecting specialized intestinal...

Magnification endoscopy after contrast enhancement with acetic acid or staining with methylene blue has been reported to be highly accurate in predicting specialized intestinal metaplasia (SIM) in Barrett's esophagus. So far, however, no data have been published on the interobserver and intra-observer variability of these new methods.

Fifty-one patients with reflux symptoms were prospectively evaluated. Endoscopy was carried out with a magnification endoscope, and video sequences were recorded in standard and zoom modes (at the 12-o'clock, 3-o'clock, 6-o'clock, and 9-o'clock positions) before and after instillation of 1.5 % acetic acid (n = 26) or staining with 0.5 % methylene blue (n = 25). Biopsies were obtained from the same locations for histopathological examination. The 102 video sequences were shown to four experienced endoscopists in a mixed and blinded fashion. The evaluation criteria used followed the published criteria; classification was carried out according to the pit-pattern structure, methylene blue positivity, and the presence of villous structures. Finally, a general statement on suspected SIM in relation to Barrett's esophagus was requested.

With regard to the criteria selected for evaluation, there was a high level of interobserver variability among the four examiners (all kappa < 0.4). SIM was histologically detectable in 60.8 % of the patients. The accuracy of all of the examiners for predicting SIM by magnification endoscopy was around 50 %, with no differences observed before and after instillation of acetic acid or methylene blue staining.

The suggested criteria for identifying SIM using magnification endoscopy are associated with a high level of interobserver variability. When evaluated in a blinded manner, staining techniques do not significantly improve the yield for detecting SIM at the esophagogastric junction.