The main purpose of endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is in the staging workup of patients with lung cancer. However, the indications for EBUS-TBNA have widened to include the diagnosis of mediastinal lymphadenopathy. A biopsy via mediastinoscopy has been the standard diagnostic procedure for mediastinal lymphadenopathy of unknown etiology. However, mediastinoscopy is invasive and has associated morbidity.

For isolated mediastinal lymphadenopathy, EBUS-TBNA has shown promising diagnostic results. EBUS-TBNA prevented the need for 87% of mediastinoscopies, and its sensitivity is 92% [1]. In a previous study on 101 patients with mediastinal and hilar lymphadenopathy or suspected lung cancer seen prospectively, EBUS-TBNA had a sensitivity of 95.1% for the correct diagnosis [2]. Choi et al. [3] studied 56 patients who underwent EBUS-TBNA as an initial diagnostic tool for enlarged lymph nodes. All diagnostic accuracy of EBUS-TBNA regardless of procedure purpose was calculated to be 83.9%. Overall, that study's diagnostic yield was comparable to other previous studies, although the diagnostic yield of benign disease was much lower than in previous studies.

Recent studies have suggested that EBUS-TBNA is useful for the diagnosis not only malignant lymphadenopathy but also benign lesions, such as sarcoidosis or tuberculosis lymphadenopathy. Of 65 patients with suspected sarcoidosis, Wong et al. [4] diagnosed sarcoidosis in 61 (93.8%) using EBUS-TBNA. Garwood et al. [5] found that EBUS-TBNA had a sensitivity of 85% for the primary diagnosis of sarcoidosis. Navani et al. [6] found that EBUS-TBNA was diagnostic in 94% of 156 patients with tuberculous intrathoracic lymphadenopathy. Moreover, EBUS-TBNA provided a positive culture in 47% and identified eight cases of drug-resistant tuberculosis. In contrast, in Choi et al. [3], the diagnostic yield of EBUS-TBNA was 50% for tuberculosis and 60% for sarcoidosis; the reason for the low diagnostic yield remains obscure, and needs further study.

We should be cautious when using EBUS-TBNA as the initial diagnostic tool in patients with lymphoma because of the relatively low diagnostic yield. Steinfort et al. [7] reviewed a prospectively recorded database of consecutive patients with suspected lymphoma who underwent EBUS-TBNA. The sensitivity of EBUS-TBNA for detecting lymphoma was 76%, and that for a definitive diagnosis of lymphoma was 57%. One possible explanation for this low diagnostic yield for lymphoma was the relatively small volume of the EBUS-TBNA samples [6]. Some subtypes of lymphoma, such as marginal zone and follicular lymphomas, are difficult to diagnose definitively from small secimens.

In conclusion, EBUS-TBNA can be an initial diagnostic tool for mediastinal lymphadenopathy. If does not provide a diagnosis, mediastinoscopy can be considered the next step.