To address these issues, we conducted a prospective study to describe the learning curve of an experienced bronchoscopy team in a unselected sample of patients with mediastinal or hilar lymph nodes. Team performance was analyzed not only in terms of the diagnostic yield of the procedure, but also to evaluate procedure length, number of lymph node passes performed to obtain adequate samples, and number of lymph nodes studied per patient. We included unselected consecutive patients with mediastinal or hilar lymph node >10 mm in the short axis on CT scan or >5 mm in the presence of PET scan-positive uptake. The study involved a team of two bronchoscopists with >10 years experience in diagnostic bronchoscopy, including conventional transbronchial needle biopsy. The bronchoscopists had previously conducted extensive theoretical training and several practical workshops and performed between five and 10 supervised procedures. The learning curve was evaluated by analyzing consecutive groups of 20 patients, the number of adequate samples obtained, and the diagnostic accuracy of the procedure. We also recorded the other variables shown in Table 1. Over a period of 13 months, EBUS-TBNA was performed for 215 lymph nodes in 120 patients (71 with lung cancer, 16 with extrapulmonary carcinomas, seven with lymphomas, and 26 with several nonmalignant pathologies). Table 1 shows these values in groups of 20 consecutive patients. No serious complications were reported.