Abstract
BACKGROUND: Robotic-assisted bronchoscopy (RAB) is an emerging diagnostic tool that combines robotics and bronchoscopy. This meta-analysis aimed to comprehensively evaluate the performance and safety of RAB for pulmonary nodule diagnosis. METHODS: PubMed, Embase, Cochrane Library, and Web of Science were searched from their inception up to 4 November 2024. The quality of the studies was assessed by the Quality Assessment of Diagnostic Accuracy Studies-2. Random and fixed effects models were used to estimate the pooled diagnosis yield in strict or intermediate criteria, sensitivity for malignancy, and complication rate of RAB in pulmonary nodule diagnosis, with rates (%) and 95% confidence intervals (CIs). RESULTS: In total, 27 cohort studies were included. The pooled diagnostic yields of RAB for pulmonary nodules were 69.6% (95%CI: 61.8%-76.8%) for strict criteria and 86.6% (95%CI: 83.7%-89.2%) for intermediate criteria, with a sensitivity for malignancy of 85.4% (95%CI: 83.0%-87.7%). The pooled complication rate was estimated to be 3.0% (total pneumothorax, 2.0%; pneumothorax that required intervention, 0.5%; bleeding, 0.1%). The diagnostic yields were different ( P < 0.05) among subgroups of patients based on total number of biopsies (≤100 vs. >100; 83.6% vs. 69.6%), prevalence for malignancy (<60% vs. ≥60%; 66.6% vs. 83.1%), radial endobronchial ultrasound view (concentric vs. eccentric vs. invisible; 88.6% vs. 84.5% vs. 46.0%). A difference ( P = 0.005) in sensitivity for malignancy was observed between the group with average lesion sizes ≤20 mm and the group with sizes >20 mm (86.4% vs. 77.5%). CONCLUSION: RAB may be effective and safe in pulmonary nodule diagnosis, offering promising prospects for clinical application. The heterogeneity of diagnostic yield may be driven by different diagnostic criteria. Moreover, the current studies of RAB in pulmonary nodule diagnosis are single-arm studies, and more large-scale randomized controlled trials are needed.