Comparison of autofluorescence and white-light bronchoscopies performed with the Evis Lucera Spectrum for the detection of bronchial cancers: a meta-analysis

Abstract

Background: Many recent studies have reported that autofluorescence bronchoscopy (AFB) has a superior sensitivity and decreased specificity in the diagnosis of bronchial cancers when compared with white-light bronchoscopy (WLB). We specifically analyzed the diagnostic performances of autofluorescence imaging video bronchoscopy (AFI) performed with the Evis Lucera Spectrum from Olympus, which is a relatively novel approach in detecting and delineating bronchial cancers, and compared it to the older WLB method.

Methods: We searched the PubMed, Embase, Web of Science, and CNKI databases from inception to July 12^th, 2018 for trials in which patients were diagnosed with lung cancer via concurrent or combined use of AFI and WLB. The included studies were required to have a histologic diagnosis as the gold standard comparison, and a sufficient amount of data was extracted to assess the diagnostic capacity. A 2×2 table was constructed, and the area under the receiver-operating characteristic curve (AUC) of AFI and WLB was estimated by using a stochastic model for diagnostic meta-analysis using STATA software.

Results: A total of 10 articles were eligible for the meta analysis, comprising 1,830 patients with complete data included in the analysis. AFI showed a superior sensitivity of 0.92 (95% CI, 0.88-0.95) over WLB's 0.70 (95% CI, 0.58-0.80) with P<0.01, and a comparable specificity of 0.67 (95% CI, 0.51-0.80) compared with WLB's 0.78 (95% CI, 0.68-0.86) with P=0.056. Egger's test P value (0.225) demonstrated that there was no publication bias.

Conclusions: Our research showed that in the evaluation of bronchial cancers, AFI was superior to conventional WLB. With its higher sensitivity, AFI could be valuable for avoiding misdiagnosis.

Keywords: Autofluorescence bronchoscopy (AFB); Evis Lucera Spectrum; bronchial cancer; meta-analysis; white-light bronchoscopy (WLB).

Figure 1

Study process screening. We carried out a systematic search of the PubMed, Embase, Web of Science and CNKI databases. The inclusion and exclusion criteria for studies are mentioned in the text. The reasons for exclusion are visible in the figure.

Figure 2

Quality assessment of the studies. Quality assessment of these studies was performed using the Cochrane Collaboration’s risk-of-bias tool by RevMan 5.3.3 software.

Figure 3

Forest plot showing study-specific (right-axis) and mean sensitivity along with specificity of AFI, with corresponding heterogeneity statistics. The sensitivity of AFI ranged from 0.65 to 0.98, with an I² of 72.29 (range, 54.59–90), while the specificity varied ranged 0.21 to 0.92, with an I² of 95.10 (range, 93.17–97.02), as determined by the STATA 14 software. The pooled sensitivity and specificity of AFI were 0.92 (95% confidence interval, 0.88–0.95) and 0.67 (95% confidence interval, 0.51–0.80). AFI, autofluorescence imaging video bronchoscopy.

Figure 4

Forest plot showing study-specific (right-axis) and mean sensitivity along with the specificity of WLB, with corresponding heterogeneity statistics. The sensitivity of WLB ranged from 0.18 to 0.94, with an I² of 87.37 (range, 80.82–93.92), and the specificity ranged from 0.50 to 0.92, with an I² of 88.56 (range, 82.79–94.33). The pooled sensitivity and specificity of WLB were 0.70 (95% confidence interval, 0.58–0.80) and 0.78 (95% confidence interval, 0.68–0.86). WLB, white-light bronchoscopy.

Figure 5

Summary ROC curve with confidence and prediction contour around mean operating sensitivity and specificity point of AFI. The AUC of AFI was 0.92 (range, 0.89–0.94). AFI, autofluorescence imaging video bronchoscopy; ROC, receiver operating characteristic; AUC, the area under the ROC curve.

Figure 6

Summary ROC curve with confidence and prediction contour around the mean operating sensitivity and specificity point of WLB. The AUC of WLB was 0.81 (range, 0.77–0.84). WLB, white-light bronchoscopy; ROC, receiver operating characteristic; AUC, the area under the ROC curve.

Figure 7

Sensitivity analysis was used to assess the stability of the results; the 95% confidence intervals of each trial overlap with each other, proving our eligible stability.