Narrow-band imaging for screening of oral premalignant or cancerous lesions: A systematic review and meta-analysis.

BACKGROUND

It remains unclear whether the use of adjunctive diagnostic screening methods improves the diagnostic efficacies of oral premalignant and cancerous lesions.

OBJECTIVE OF REVIEW

We evaluated the diagnostic accuracy of narrow-band imaging used to detect oral cancer and precancerous lesions defined employing different narrow-band imaging criteria.

TYPE OF REVIEW

Systematic review and meta-analyses.

SEARCH STRATEGY

We searched PubMed, Scopus, the Web of Science, Embase, Google Scholar and the Cochrane Central Register of Controlled Trials to May 2020.

EVALUATION METHODS

Three different criteria for oral mucosal vascular changes using narrow-band imaging were compared: class I: well-demarcated brownish areas with thick dark spots and/or winding vessels; class II: intraepithelial papillary capillary looping of grades 2, 3 and 4; and class III: intraepithelial papillary capillary looping of grades 3 and 4. Methodological quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies (ver. 2) tool. We compared narrow-band imaging to conventional white-light imaging.

RESULTS

We included 10 prospective or retrospective studies (1374 patients). To detect all dysplastic and cancerous lesions, the class I criteria afforded the optimal specificity and sensitivity; the area under the summary receiver operating characteristic curve was 0.918. To detect highly dysplastic and advanced cancerous lesions, the class III criteria afforded appropriate specificity and sensitivity. The summary receiver operating characteristic curve was 0.905. When using the class III criteria, narrow-band imaging afforded better specificity (0.941 [range 0.920, 0.9572], P < .0001) compared to white-light imaging (0.520 [range 0.409, 0.629]). However, the white-light imaging data were inconsistent and the ranges were broad; narrow-band imaging may be considerably more accurate than white-light imaging when using the class III criteria.

CONCLUSION

Narrow-band imaging diagnosed oral premalignant or cancerous lesions much more reliably than white-light imaging.