Twenty-one studies, with 3,266 participants (range 30 to 1,173) and 8,851 biopsies (range 77 to 2,907), were included in the review. The authors did not report the results of the quality assessment.
Intra-epithelial neoplasia: The pooled sensitivity estimates were 84.6% using autofluorescence and white-light bronchoscopy and 42.5% using white-light bronchoscopy alone (19 studies). The combination was significantly more sensitive than white-light bronchoscopy alone (RR 2.04, 95% CI 1.72 to 2.42); there was significant between-study heterogeneity. Subgroup analyses indicated that Onco-LIFE for autofluorescence bronchoscopy resulted in the highest relative risk (4.25, 95% CI 1.56 to 11.55), but only one study used this device.
Invasive cancer: The pooled sensitivity estimates were 94.7% with the combination and 88.5% with white-light bronchoscopy alone (14 studies). The combination was significantly more sensitive (RR 1.15, 95% CI 1.05 to 1.26); there was no evidence of significant heterogeneity. Subgroup analyses indicated that, when devices were considered separately, only autofluorescence bronchoscopy using LIFE resulted in a statistically significant relative risk (1.21, 95% CI 1.04 to 1.40).
Any cancer: The pooled overall estimates of specificity were 60.9% for the combination and 79.7% for white-light bronchoscopy alone (16 studies). The combination was significantly less specific (RR 0.65, 95% CI 0.59 to 0.73); there was significant heterogeneity. Subgroup analyses indicated that the difference in specificity was greatest (lowest RR) when LIFE was used for autofluorescence bronchoscopy (RR 0.56, 95% CI 0.47 to 0.68).
There was evidence of publication bias.