Histological assessment & use of immunohistochemical markers for detection of dysplasia in Barrett's esophageal mucosa.

BACKGROUND

Histological assessment of dysplasia in Barrett's esophagus (BE) has high inter-observer variability. Hence, use of ancillary markers for early detection of dysplasia in BE is an important clinical question.

METHODS

In this retrospective study consecutive cases of BE (n = 59), over a period of 4 years were included. Hematoxylin and eosin stained sections were reviewed independently by 3 senior qualified pathologists, who graded the dysplasia according to the Vienna Classification system and inter-observer agreement was analysed using the Kappa statistics. Subsequently Alpha-Methyl Acyl-CoA Racemase (AMACR), p53, CyclinD1, β-catenin, H2AX and M30 immunohistochemical (IHC) stains were examined on the following disease categories: BE with no dysplasia [NFD] (45), BE with indefinite for dysplasia (IFD) (4), low grade dysplasia (LGD) (3), high grade dysplasia (HGD) (2) and in adenocarcinomas (5). H score was calculated by adding up products of different grades of stain distribution and stain intensities (range of scores 0-300).

RESULTS

Among the 3 pathologists, overall agreement was poor (k 0.06; 95% CI -0.089 to 0.145), with highest disagreement noted for differentiating the LGD and IFDs (k = 0.21). After revising the histological criteria, the kappa improved to 0.53. Among the IHC stains performed, p53, β-catenin, H2AX and M30 stains were significantly useful to differentiate between IFD and LGD (P values: 0.04, 0.004, 0.05 & 0.04, respectively). AMACR and β-catenin stains though were up-regulated in HGD/adenocarcinomas than in other categories, their expression were not statistically different between the IFD and LGDs.

CONCLUSIONS

A detail histological scoring system may bring uniformity in histological interpretation of dysplasia in BE. Using a combined panel of IHC stains seems helpful in detection of dysplasia in BE, especially to differentiate the IFD and LGD changes in BE.