作者: Mário Dinis-Ribeiro , Altamiro da Costa-Pereira , Carlos Lopes , Lúcio Lara-Santos , Mateus Guilherme
DOI: 10.1067/MGE.2003.145
关键词:
摘要: Background: The aim of this study was to define the reproducibility and accuracy magnification chromoendoscopy for diagnosis lesions associated with gastric cancer (intestinal metaplasia dysplasia). Methods: A total 136 patients previously diagnosed 5 gastrectomy specimens were studied. Endoscopic examination performed a endoscope after methylene blue (1%) spraying. According differences in color mucosal pattern, groups subgroups endoscopic images defined, biopsies taken (n=462). Five endoscopists asked classify individually 2 per subgroup on separate occasions. Results: Three defined: nonmetaplastic, nondysplastic mucosa (I); metaplastic (II); dysplastic (III). Ten defined according pit pattern: round small (IA), tubular (IB), coarse (IC), course pits straight (ID); irregular marks (IIA), (IIB), villi (IIC), (IID); loss clear depression (IIIA) or slight elevation (IIIB). kappa statistic intraobserver agreement classification 0.86; interobserver agreement, it 0.74. For into subgroups, values ranged from 0.48 0.78. 85% areas classified endoscopically as Group I (n=146), no gastritis described at histologic examination; 83% those II (n=198), intestinal found. Subgroups IIA IIB more often complete (62%), IIC IID incomplete (67%); III (n=118), dysplasia histopathologically 33%. dysplasia, specificity 81% (95% CI [77%, 85%]) negative predictive value 99% [99%, 100%]). Conclusions: Gastric patterns seem reproducible valid cancer. This procedure may improve follow-up individuals high-risk cancer, least exclusion severe lesions.