[Detection of MG7 antigen expression in predicting high-risk in gastric cancer by automatic image analysis]

Zhonghua Yu Fang Yi Xue Za Zhi. 1996 Sep;30(5):286-8.
[Article in Chinese]

Abstract

To study the predictive value of gastric cancer monoclonal antigen (MG7-Ag) in detecting atypical hyperplasia of gastric mucosa, its expression in 72 cases of the illness with follow-up was determined quantitatively by computer-aided image analysis system with immunohistochemical stain for MG7 antibody. Critical value of MG7-Ag expression for high-risk in gastric cancer was explored preliminarily by Youden index and receiver operator characteristic curve. Results of immunohistochemical stain showed there were 24 with positive expression (70.6%) in 34 cancerous cases, and only seven in 38 non-cancerous cases (18.4%), with a significant difference. Image analysis showed an average MG7-Ag density equal to or greater than 0.19 was regarded as the critical value for high-risk in gastric cancer, which could be used to discriminate atypical hyperplasia from canceration, with a sensitivity of 82.4%, a specificity of 78.9%, and an accuracy of 80.6%. Thus, it could be concluded that MG7-Ag expression correlated closely with canceration of gastric atypical hyperplasia, positive MG7-Ag expression in gastric mucosa of patients with atypical hyperplasia, especially in cases with a density of equal to or greater than 0.19, was an indicator for high risk in canceration. If they are put under strict follow-up, early detection of gastric cancer will be improved.

Publication types

  • English Abstract
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Antibodies, Monoclonal
  • Antibodies, Neoplasm / analysis*
  • Follow-Up Studies
  • Gastric Mucosa / pathology
  • Humans
  • Hyperplasia
  • Image Processing, Computer-Assisted
  • Precancerous Conditions / immunology*
  • Precancerous Conditions / pathology
  • Risk Factors
  • Stomach Neoplasms / immunology*
  • Stomach Neoplasms / pathology

Substances

  • Antibodies, Monoclonal
  • Antibodies, Neoplasm