Computed Tomography Bronchus Sign and the Diagnostic Yield of Guided Bronchoscopy for Peripheral Pulmonary Lesions. A Systematic Review and Meta-Analysis.

RATIONALE: Indeterminate peripheral pulmonary lesions (PPLs) often require tissue diagnosis. If nonsurgical biopsy techniques are considered, deciding between bronchoscopic transbronchial versus computed tomography-guided transthoracic biopsy can be difficult. The former has a low diagnostic yield with a low complication risk, whereas the latter has a better diagnostic yield but a higher complication rate. Investigators have looked at various lesion characteristics that can predict the diagnostic yield of guided bronchoscopic biopsies. Although consensus exists that larger size and proximity to the hilum increase the diagnostic yield, there is ongoing debate about the association between computed tomography bronchus sign (air-filled bronchus in close proximity of the lesion as seen on computed tomography imaging) and the diagnostic yield of guided bronchoscopic modalities.

OBJECTIVES: To perform a meta-analysis and systematic review, determining the association between computed tomography bronchus sign and the diagnostic yield of guided bronchoscopy for PPLs.

METHODS: MEDLINE, Embase, Scopus, and Google Scholar were searched in January 2018 for guided bronchoscopy studies that had assessed the impact of computed tomography bronchus sign on the diagnostic yield. The quality of included studies was assessed using Quality Assessment of Diagnostic Accuracy Studies-2 tool. Meta-analysis was performed using MedCalc (version 18). Odds ratios were used to compare yield of lesions with and without bronchus sign. Random effects model was used when significant heterogeneity was observed (I2  > 40%).

RESULTS: For 2,199 lesions with computed tomography bronchus sign, the overall weighted diagnostic yield was 74.1% (95% confidence interval, 68.3-79.5%). For 971 lesions without computed tomography bronchus sign, the overall weighted diagnostic yield was 49.6% (95% confidence interval, 39.6-59.5%). The odds ratio for successfully diagnosing a lesion with computed tomography bronchus sign was 3.4 (95% confidence interval, 2.4-5.0). Possible sources of heterogeneity in the meta-analysis included differences in study designs, guidance modalities, and cancer prevalence. The odds ratio for successfully diagnosing a lesion with computed tomography bronchus sign was relatively lower for prospective studies.

CONCLUSIONS: PPLs with computed tomography bronchus sign are more likely to be diagnosed with guided bronchoscopy than the lesions without computed tomography bronchus sign. Clinicians should consider this, along with the lesion size and distance from the hilum, when contemplating guided bronchoscopy for PPLs.