On the off chance you someday need to diagnose tuberculosis from intrathoracic lymphadenopathy (I suppose in the rare patient with smear-negative sputum, which is more common in HIV co-infection), endobronchial ultrasound can help you out, say Navani et al.
They report on 156 consecutive patients (over 2 years at 4 centers) found to have tuberculous lymph nodes in their chests. In 146 (94%), EBUS-TBNA made the diagnosis. (Of the 10 others, 4 underwent mediastinoscopy which made the diagnosis; the other 6 had a clear response to antituberculous drugs).
Only half of the EBUS-TBNA samples were culture-positive, but 86% had tuberculi on pathology.
TBNA shouldn't aerosolize tuberculosis; authors reported no transmission to medical care team members. They recommend using EBUS-TBNA is "a safe and effective first-line investigation in patients with tuberculous intrathoracic lymphadenopathy." But isn't the point that you don't know they have TB prior to biopsying them?
The question is important, because it gets at whether routine EBUS-TBNA is indicated for asymptomatic thoracic lymphadenopathy. Fully 22% of their patients were described as asymptomatic (more than half had cough, fever, weight loss and/or night sweats), and only 11% were HIV(+).
This contrasts interestingly with a recent paper that showed the diagnostic yield of EBUS TBNA for asymptomatic mediastinal lymphadenopathy was close to nil, except in diagnosing occult sarcoidosis. Those authors found exactly zero cases of TB. It's probably all about your population and your pretest probability for tuberculosis.
Navani N et al. Utility of endobronchial ultrasound-guided transbronchial needle aspiration in patients with tuberculous intrathoracic lymphadenopathy: a multicentre study. Thorax 2011;66:889-893.