Diagnosis of gastric intestinal metaplasia with confocal laser endomicroscopy in vivo: a prospective study

Endoscopy. 2008 Jul;40(7):547-53. doi: 10.1055/s-2007-995633.


Background and study aims: Gastric intestinal metaplasia (GIM) is a risk factor for development of intestinal-type gastric cancer. We aimed to assess the usefulness of confocal laser endomicroscopy (CLE) in diagnosing GIM.

Patients and methods: 28 patients with known GIM underwent CLE, and CLE criteria for diagnosis of GIM were developed. In addition, 53 consecutive patients with known or suspected GIM were prospectively evaluated.

Results: GIM was identified if any of the following three features were present in an image field: goblet cells, columnar absorptive cells and brush border, and villiform foveolar epithelium. GIM was then classified as complete or incomplete according to the shape of the goblet cells, the presence of absorptive cells or brush border, and the architecture of vessels and crypts. In a prospective study, a total of 13 670 CLE images were obtained. Among 267 sites from 53 patients, 160 from 36 patients were diagnosed histopathologically as GIM. The sensitivities of conventional endoscopy and CLE for GIM were 36.88 % vs. 98.13 %, and the specificities were 91.59 % vs. 95.33 %, respectively. The kappa value for the correlation with histological findings was 0.25 for conventional endoscopy vs. 0.94 for CLE. The sensitivity and specificity of CLE were 68.03 % and 89.66 %, respectively, for the diagnosis of complete GIM, and 68.42 % and 83.41 %, respectively, for incomplete GIM; the kappa value for the correspondence between CLE and histological findings was 0.67.

Conclusion: CLE is a useful and potentially important method for the diagnosis and classification of GIM in vivo.

MeSH terms

  • Adult
  • Aged
  • Biopsy
  • Female
  • Gastric Mucosa / pathology*
  • Goblet Cells / pathology
  • Humans
  • Male
  • Microscopy, Confocal* / methods
  • Microvilli / pathology
  • Middle Aged
  • Prospective Studies
  • Sensitivity and Specificity