BACKGROUND: Transesophageal introduction of the endobronchial ultrasound (EBUS) videobronchoscope allows pulmonologists to perform endoscopic ultrasound fine-needle aspiration (EUS-B-FNA) of mediastinal lesions. Safety, diagnostic accuracy, and feasibility of EUS-B-FNA in evaluation of pulmonary parenchymal lesions are not established. METHODS: All patients undergoing pulmonologist-performed EUS-B-FNA of parenchymal lung lesions at 2 tertiary centers were included in this prospective observational cohort study. RESULTS: EUS-B-FNA sampling of parenchymal lesions was performed in 27 patients. Mean (±SD) lesion size was 36±16 mm. Seven lesions were ≤18 mm. Pneumothorax occurred in 1 patient (3.7%, 95% confidence interval, 0.001%-19%). Ten target lesions (36%) were in locations inaccessible to bronchoscopic sampling via the airways, and 9 lesions were inaccessible to EBUS-guided transbronchial needle aspiration and in locations associated with low diagnostic yield from radial EBUS. EUS-B-FNA was diagnostic in 26 patients (96%), and sensitivity of EUS-B-FNA was 100% (95% confidence interval, 87%-100%) for both lung cancer (n=21) and for pulmonary metastatic lesions (n=5). CONCLUSIONS: Pulmonologist-performed EUS-B-FNA is safe and accurate in the evaluation parenchymal lung lesions. Diagnostic accuracy is high. EUS-B-FNA may achieve access to sites not amenable to other forms of bronchoscopic sampling, or increase diagnostic accuracy in patients where anatomic position predicts a low diagnostic yield.