Background: Electromagnetic navigation bronchoscopy (ENB) is an emerging endoscopic technique for the diagnosis of peripheral lung lesions. A thorough analysis of ENB's yield and safety is required for comparison to other sampling modalities. Objectives: To describe ENB's yield and safety profile. Methods: The MEDLINE and EMBASE databases were systematically searched for studies reporting ENB's yield for peripheral lung lesions. Two independent investigators extracted data and rated each study on a scale of methodological quality. Clearly defined performance outcomes were reconstructed and meta-analyzed. Subgroup analysis and meta-regression were used to identify possible sources of study heterogeneity. Results: A total of 15 trials were included (1,033 lung nodules). A positive and definitive diagnosis was obtained after 64.9% of all ENB procedures (95% CI 59.2-70.3). Overall diagnostic accuracy was 73.9% (95% CI 68.0-79.2). Sensitivity to detect cancer was 71.1% (95% CI 64.6-76.8), with a negative predictive value of 52.1% (95% CI 43.5-60.6). Pneumothorax occurred in 3.1% of patients, requiring chest tube drainage in 1.6% of these cases. Original trials identified 6 variables associated with higher ENB yields: nodule location in the upper or middle lobes, nodule size, lower registration error, presence of a bronchus sign on CT imaging, combined use of an ultrasonic radial probe, and catheter suctioning as a sampling technique. Heterogeneity exploration revealed that studies using general anesthesia or rapid on-site cytological evaluation reported better yields. Conclusions: ENB is effective and particularly safe. Prospective studies are needed to clarify the role of several variables conditioning the yield of this technique.