Endobronchial ultrasound-guided transbronchial needle aspiration of mediastinal lymphadenopathy: effect of the learning curve.

This study aimed to evaluate the learning curve and efficacy of endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) in the examination of mediastinal or hilar lymph nodes suspected of having cancer or of benign disease such as sarcoidosis. A success-adjusted cumulative sum model was used to evaluate the learning curve for diagnostic rates and operation time. A total of 99 patients (77 men and 22 women) who underwent EBUS-TBNA from April 2011 to March 2012 in a single centre were analysed retrospectively. The quantity of lymph node sampling was deemed to be appropriate for histopathological examination in 97 of 99 patients (97%). Twenty-three cases (23%) were clearly diagnosed with neoplastic disease, 60 (60%) with reactive hyperplasia, 11 (11%) with granulomatosis and 3 (3%) histopathologically suspicious for lymph node metastasis. The sensitivity, specificity, and positive and negative predictive values and diagnostic accuracy for EBUS-TBNA were 80, 100, 100, 87.1 and 91.5%, respectively. According to the learning curve analysis, the ability to perform EBUS required performing approximately 37 procedures for the trials. In conclusion, more successful results are obtained after a certain learning curve, as is the case for every other invasive procedure.