Abstract
BACKGROUND: Gastric intestinal metaplasia (GIM) can be missed by random gastric biopsies taken during white light endoscopy. Narrow band imaging (NBI) may potentially improve the detection of GIM. However, pooled estimates from prospective studies are lacking and the diagnostic accuracy of NBI in detecting GIM needs to be more precisely defined. The aim of this systematic review and meta-analysis was to study the diagnostic performance of NBI in detecting GIM. METHODS: PubMed/Medline and EMBASE were screened for studies examining GIM in relation to NBI. Data from each study were extracted and calculations of pooled sensitivity, specificity, likelihood ratios, diagnostic odds ratios (DORs), and areas under the curve (AUCs) were performed. Fixed or random effects models, were used as appropriate, depending on the presence of significant heterogeneity. RESULTS: We included 11 eligible studies in the meta-analysis, comprising 1672 patients. NBI showed a pooled sensitivity of 80% (95% confidence interval [CI] 69-87), specificity of 93% (95%CI 85-97), DOR 48 (95%CI 20-121), and AUC of 0.93 (95%CI 0.91-0.95) in detecting GIM. CONCLUSIONS: This meta-analysis showed that NBI is a reliable endoscopic means of detecting GIM. NBI with magnification performed better than NBI without magnification. However, better designed prospective studies are needed to precisely determine the diagnostic role of NBI, especially in high-risk populations where early detection of GIM can impact gastric cancer prevention and survival.