Background/aims: To testify proposed endoscopic criteria in surveying intestinal metaplasia (IM), we conducted a prospective study to assess the accuracy of endoscopic diagnosis of IM correlated with histology.
Methodology: Between January 1996 and July 1997, 87 subjects undergoing endoscopic examination as part of a health check-up were enrolled. Subjects whose endoscopic features showed whitish color change with plaques, patches or homogeneous discoloration on gastric mucosa were diagnosed as IM and enrolled in group A, while subjects without any of these endoscopic features were enrolled in group B. Endoscopic biopsies were taken from areas with whitish changes and normal-looking mucosae for histological correlation.
Results: There were 8 subjects with whitish plaques, 29 subjects with whitish patches, and 8 subjects with homogeneous whitish discoloration enrolled in group A. In contrast, 42 subjects without any endoscopic features of IM were enrolled in group B. After verification by endoscopic biopsy with histologic examination, 30 and 10 subjects were histologically confirmed to have IM in groups A and B, respectively. The overall sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of endoscopic diagnosis was 75%, 68.1%, 66.7%, 76.2%, and 71.3%, respectively. There was a positive relationship between age and sensitivity, and tendency of an inverse relationship between age and specificity.
Conclusions: The proposed endoscopic criteria are valuable in detecting three-fourths of IM patients during routine endoscopic examination. It was more effective to detect IM in elderly patients and to exclude IM in younger subjects by these criteria.