Narrow-band imaging bronchoscopy in the detection of premalignant airway lesions: a meta-analysis of diagnostic test accuracy

Objectives: Both autofluorescence imaging bronchoscopy and narrow-band imaging have shown promise in the detection of premalignant airway lesions, each by utilizing different bandwidths of lights for better characterization of the mucosal and submucosal vascular grid. Since previously published meta-analyses have shown poor specificity of autofluorescence imaging bronchoscopy, we specifically studied the diagnostic accuracy of narrow-band imaging alone and in combination with autofluorescence imaging bronchoscopy in the detection of premalignant airway lesions. Methods: After an extensive search of eligible studies from PubMed and Medline, extracted data were pooled with weighted averages. Symmetrical summary–receiver operating characteristic curves were constructed to summarize the results quantitatively. Study heterogeneity was assessed by the I 2 index. Results: Analysis of data from eight studies on narrow-band imaging showed a pooled sensitivity of 0.80 [95% confidence interval (CI): 0.77–0.83] and a pooled specificity of 0.84 (95% CI: 0.81–0.86). Summary–receiver operating characteristic curves from the data on narrow-band imaging calculated an area-under-the-curve of 0.908 (standard error 0.01). The diagnostic odds ratio of narrow-band imaging was 31.49 (95% CI: 12.17–81.45). Data from studies where narrow-band imaging and autofluorescence imaging bronchoscopy were used together showed a pooled sensitivity, specificity, area-under-the-curve and diagnostic odds ratios of 0.86 (95% CI: 0.82–0.89), 0.75 (95% CI: 0.71–0.79), 0.964 (standard error 0.05) and 27.96 (95% CI: 3.04–257.21), respectively. Conclusions: Our findings indicate that in the evaluation of premalignant airway lesions, narrow-band imaging has a higher sensitivity, specificity and diagnostic odds ratios compared with autofluorescence imaging bronchoscopy. However, combining autofluorescence imaging bronchoscopy and narrow-band imaging does not significantly improve test performance characteristics.