Diagnostic performance of chromoendoscopy and narrow band imaging for colonic neoplasms: a meta-analysis

Aim: We conducted a meta-analysis to compare the diagnostic test performance of chromoendoscopy and narrow band imaging (NBI) for colonic neoplasms.

Method: MEDLINE, EMBASE and the Cochrane Library were searched (1966 to March 2009). Articles were included if: (i) chromoendoscopy or NBI was used, (ii) sensitivity and specificity were reported; (iii) absolute numbers of true-positive, false-positive, true-negative and false-negative results were provided or could be calculated; and (iv) pathology was used as the reference standard. Sensitivity and specificity were pooled using random effects model. Secondary analyses were conducted by limiting the studies in which magnifying endoscopy was used alone as a diagnostic modality, and polyp size and macroscopic appearance of lesions were not considered.

Results: Of 1342 screened articles, 27 met the inclusion criteria. Pooled sensitivity for chromoendoscopy and NBI was 0.94 (95% CI, 0.92-0.95) and 0.94 (0.91-0.97), and specificity was 0.82 (0.77-0.88) and 0.86 (0.83-0.89), respectively. There were no differences in sensitivity (P = 0.99) or specificity (P = 0.54) between the two methods. In the secondary analysis, pooled sensitivity for choromoendoscopy and NBI was 0.93 (95% CI, 0.90-0.97) and 0.96 (0.93-0.99) and specificity was 0.80 (0.73-0.87) and 0.85 (0.78-0.92). respectively. Overall, the pooled false-negative rate was 0.057 (95% CI, 0.040-0.73) for chromoendoscopy and 0.057 (95% CI, 0.028-0.085) for NBI.

Conclusion: Chromoendoscopy and NBI had similar diagnostic test characteristics in the assessment of colonic neoplasms; however, the false-negative rate for both methods of 5.7% is an unacceptably high rate and currently therefore, neither method is ready for general use.