[Prediction of the outcome of dysplasia of esophageal epithelium by high resolution image analysis]

Zhonghua Zhong Liu Za Zhi. 1999 Nov;21(6):439-43.
[Article in Chinese]

Abstract

Objective: To predict the outcome of dysplasia of esophageal epithelium by means of high resolution image analysis(HRIA).

Methods: Asymptomatic adults were examined for balloon cytology of the esophagus in 1983 from Heshun Commune of Linxian County. Ninety three cases of severe dysplasia and 122 cases of mild dysplasia of the esophagus were selected for this study. By means of an Axiomat-microscope equipped with TV-camera, 100 normal nuclei of well-preserved cells in the intermediate layer of Pap-stained squamous epithelium were randomly examined.

Results: Of the 93 cytologically diagnosed severe dysplasia cases, 24, 14 and 7 progressed to carcinoma in 3, 5 and 9 years, respectively. In the other 48 cases, dysplasia remained stable or regressed to normal. The other cases were used as the control. According to chromatin features, correct diagnosis of cases was achieved by HRIA in 75.0%(18/24), 85.7%(12/14) and 85.7%(6/7) of the cases examined, respectively (P < 0.001). Of the 122 cytologically diagnosed mild dysplasia, 16, 13 and 12 cases progressed to carcinoma in 3, 5 and 9 years, respectively. The other 81 cases remained stable or regressed to normal. Correct diagnosis was made by HRIA in 93.8%(15/16), 76.9%(10/13) and 83.3%(10/12) of the cases examined, respectively (P < 0.001).

Conclusion: Chromatin nuclear features examined by HRIA can predict the outcome of precancerous lesions and discriminate progressor from non-progressor ones. It can be used as surrogate endpoint biomarkers for the evaluation of efficacy of chemoprevention trial.

Publication types

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

MeSH terms

  • Adult
  • Aged
  • Cytodiagnosis
  • Esophageal Neoplasms / diagnosis
  • Esophageal Neoplasms / pathology*
  • Humans
  • Image Processing, Computer-Assisted
  • Middle Aged
  • Precancerous Conditions / diagnosis
  • Precancerous Conditions / pathology*