Robotic-assisted Navigation Bronchoscopy: A Meta-Analysis of Diagnostic Yield and Complications.

Abstract

Robotic-assisted navigation bronchoscopy (RANB) is a novel method to biopsy lung nodules, with initial reports demonstrating excellent accuracy. We aimed to evaluate pooled estimates of diagnostic yields and complication rates with RANB by performing a meta-analysis of the available literature. We searched 3 databases, including PubMed, EmBase, and Web of Science. The resulting abstracts were reviewed by 2 investigators. Analyses were performed using random effects models, and diagnostic yield and complication rates were estimated after the Freeman-Tukey transformation. A total of 23 articles, comprising 1409 patients and 1541 nodules, were included in the final analysis. Mean ages ranged from 63.2 to 69.3 years. The average size of the nodules ranged between 5.9 and 25.0mm. Most patients (54.0% to 92.0%) had a current or prior smoking history in studies that reported them (n=8). The pooled diagnostic yield was 81.9% (12 studies, 838 nodules, 95% CI: 83.4%-91.0%), and the pooled sensitivity for malignancy was 87.6% (8 studies, 699 nodules, 95% CI: 81.3%-89.5%). The pooled incidence of pneumothorax rates was 0.60% (95% CI: 0.11%-1.35%). The pooled incidence of major bleeding was <0.01%. Diagnostic yield for patients with pulmonary nodules undergoing RANB is high, though may be impacted by the prevalence of malignancy, participant selection, and publication bias. Complication rates, including pneumothoraces and bleeding rates, appear low across all studies. If RANB is available, clinicians should consider utilizing this platform to biopsy pulmonary nodules.