Clinical decision making in Barrett's oesophagus can be supported by computerized immunoquantitation and morphometry of features associated with proliferation and differentiation

J Pathol. 1998 Feb;184(2):161-8. doi: 10.1002/(SICI)1096-9896(199802)184:2<161::AID-PATH971>3.0.CO;2-2.


Grading of dysplasia in Barrett's oesophagus has a therapeutic impact, but subjective grading is associated with substantial observer variation. Quantitative pathological methods could help to achieve a more accurate and reproducible diagnosis. In the present study, the immunoquantitation of p53 and Ki67 and the morphometric analysis of features associated with proliferation and differentiation were evaluated for this purpose. In slides of 35 oesophagectomy specimens, 73 areas that displayed either no dysplasia (ND), low-grade dysplasia (LGD), high-grade dysplasia (HGD), or intramucosal carcinoma (ImCa) were initially considered. Agreement on double blind examination by two experienced pathologists was reached in 58 areas, which were used as the 'learning set'. The 15 areas of disagreement were used as a second set. In the univariate analysis, the most significant differences in the learning set were found for Ki67, p53, stratification index (SI), mean nuclear area, and volume. Further multivariate analysis showed that for discrimination between ND and LGD, the combination of Ki67 and SI resulted in 94 per cent correctly classified areas. Likewise, for the discrimination between LGD and HGD, Ki67 and SI were the most powerful combination (again, 94 per cent of areas classified correctly). The discrimination between HGD and ImCa with any combination of the quantitative parameters never exceeded 80 per cent correct classification. The addition of p53 was of no value in improving the discrimination of ND vs. LGD, or of LGD vs. HGD. In the 15 original disagreement areas of the initial set of 73, three of the five ND/LGD areas could be uniquely classified as either ND or LGD by Ki67 and SI. Moreover, three of the four LGD/HGD disagreement areas could be uniquely classified with the combination of Ki67 and SI as either LGD or HGD. We conclude that the quantitative assessment of cytometric and morphometric features associated with proliferation and differentiation (especially Ki67 and SI) can be a valuable adjunct tool for clinical decision making in Barrett's oesophagus.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Barrett Esophagus / pathology*
  • Cell Differentiation
  • Cell Division
  • Diagnosis, Differential
  • Disease Progression
  • Esophageal Neoplasms / pathology*
  • Humans
  • Image Processing, Computer-Assisted
  • Immunoenzyme Techniques
  • Ki-67 Antigen / metabolism
  • Neoplasm Proteins / metabolism
  • Precancerous Conditions / pathology*
  • Tumor Suppressor Protein p53 / metabolism


  • Ki-67 Antigen
  • Neoplasm Proteins
  • Tumor Suppressor Protein p53