6533b7dafe1ef96bd126f598

RESEARCH PRODUCT

Chromoendoscopy in Barrett's oesophagus: is cresyl violet the magic bullet?

subject

description

The endoscopic detection of Barrett’s epithelium remains challenging even for modern endoscopy. This is mainly due to the fact that Barrett’s epithelium is often patchy and can easily be overlooked by conventional endoscopy with random biopsies. Thus, chromoendoscopy and magnifying endoscopy were introduced to facilitate diagnosis of Barrett’s epithelium and Barrett’s associated neoplasias. Methylene blue-aided chromoendoscopy was firstly introduced by Canto et al. [1]. The authors could show that methylene blue selectively stains specialised columnar epithelium, which is pathognomonic for Barrett’s epithelium. In contrast, dysplastic areas revealed no or weaker staining due to changes in the nucleus/plasma ratio in neoplastic cells. Thus, unstained areas may either represent gastric type epithelium or dysplastic areas [2]. However, methylene blue staining at high concentration (0.5–1%) was never widely accepted because of the time consuming procedure and the overall disappointing results with regard to the diagnosis of dysplastic or malignant lesions within columnar lined lower oesophagus (CLE) [3]. The introduction of magnifying endoscopes has opened the possibility to analyse the mucosal surface architecture within CLE. Besides the analysis of stained or unstained areas, graduation of pit openings during magnifying chromoendoscopy can be used to predict Barrett’s epithelium and Barrett’s associated neoplasias. In general, magnifying endoscopy should be used in conjunction with dyes or acetic acid. Using this approach Guelrud et al. have graduated and differentiated the surface architecture of Barrett’s epithelium. Using acetic acid to highlight the mucosal surface architecture due to protein interactions, they identified four different types of Barrett’s epithelium after local application of acetic acid [4]. In contrast to this study, Endo et al. [5] used diluted methylene blue solution (0.2%) to differentiate Barrett’s epithelium into five different types. Interestingly, they could show that tubular and villous surface architecture predicting specialised columnar epithelium was only associated in 50% and 60% with the presence of positive methylene blue staining, respectively. Both of the above studies analysed differences in surface architecture of gastric type epithelium and Barrett’s epithelium only in the absence dysplastic changes within the distal oesophagus. To further expand the diagnostic possibilities of chromoendoscopy towards dysplasia, Sharma et al. [6] used the combination of magnifying endoscopy and indigo carmine staining. Barrett’s epithelium was identified in 57 out of 62 (97%) patients with a ridged and villous surface-staining pattern and in 2 out of 12 (17%) cases with circular staining pattern. Low-grade dysplasia was found in 18 patients with ridged/villous pattern, whereas high-grade dysplasia was detected in six patients where an irregular and distorted pattern was observed. However, simultaneously Egger et al. [7] described the poor outcome of chromoendoscopy to predict dysplasia by prospectively comparing methylene blue-aided chromoendoscopy with autofluorescence and random biopsies. Furthermore, Meining et al. [8] tested the kappa values for all above mentioned classifications and found that all classifications show poor inter- and intraobserver agreement. Finally, chromoendoscopy thus far failed to convincingly show that a significant number of biopsies can be saved as compared to random biopsies. Taken together, these data questioned the value of chromoendoscopy for diagnosing Barrett’s epithelium and neoplasias. In this issue of Digestive and Liver Disease, Yuki et al. [9] have introduced a new dye, cresyl violet, for chromoendoscopy of Barrett’s epithelium. Using this dye in a large number of patients (1030), they propose a new classification for diagnosing Barrett’s epithelium and Barrett’s associated neoplasias. The design of the Yuki classification offers two advantages: first, the classification is simple and only two classes (open and closed pits) of pits are differentiated; second, the classification can be used for all entities of columnar lined lower oesophagus. Using the Yuki classification, prediction of dysplasia and Barrett’s epithelium was possible with a sensitivity of 96.0% (81.9%), specificity of 66% (95.6%) and accuracy rate of 67.4% (90.5%), respectively.