Figure 1.
Flow chart showing the process for selecting eligible studies in this meta-analysis.
Table 1.
Characteristics of the included studies.
Table 2.
The Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool for quality assessment of studies selected for the meta-analysis.
Figure 2.
Gastric intestinal metaplasia (GIM), per-patient analysis of diagnostic performance of narrow band imaging (NBI).
A. Pooled sensitivity for NBI to differentiate GIM; B. Pooled specificity for NBI to differentiate GIM; C. The summary receiver operating characteristic (SROC) cure for diagnosis by NBI. CI, confidence interval; df, degrees of freedom; AUC, area under curve; SE, standard error.
Figure 3.
Gastric intestinal metaplasia (GIM), per-patient analysis of diagnostic performance of magnification narrow band imaging (NBI-ME).
A. Pooled sensitivity for NBI-ME to differentiate GIM; B. Pooled specificity for NBI-ME to differentiate GIM; C. The summary receiver operating characteristic (SROC) cure for diagnosis by NBI-ME. CI, confidence interval; df, degrees of freedom; AUC, area under curve; SE, standard error.
Figure 4.
Gastric intestinal metaplasia (GIM), per-lesion analysis of diagnostic performance of narrow band imaging (NBI).
A. Pooled sensitivity for NBI to differentiate GIM; B. Pooled specificity for NBI to differentiate GIM; C. The summary receiver operating characteristic (SROC) cure for diagnosis by NBI. CI, confidence interval; df, degrees of freedom; AUC, area under curve; SE, standard error.
Figure 5.
Gastric intestinal metaplasia (GIM), per-lesion analysis of diagnostic performance of magnification narrow band imaging (NBI-ME).
A. Pooled sensitivity for NBI-ME to differentiate GIM; B. Pooled specificity for NBI-ME to differentiate GIM; C. The summary receiver operating characteristic (SROC) cure for diagnosis by NBI-ME. CI, confidence interval; df, degrees of freedom; AUC, area under curve; SE, standard error.
Table 3.
Meta-regression for the potential source of heterogeneity.
Figure 6.
Funnel plot was constructed by Begg's and Egger's test to evaluate publication bias.
A. Begg's test for the studies differentiating GIM on a per-patient analysis; B. Begg's test for the studies differentiating GIM on a per-lesion analysis. Lnor, ln [(TP*TN)/(FP*FN)]; s.e.of:lnor, (1/TP)+(1/FP)+(1/FN)+(1/TN).