Bronchoscopy will not show pathological findings always in bronchial tree TB. Even broncos optic biopsy is pos in about 60 to 85% cases. Modern CT scan may be of some diagnostic use.  In these cases the diagnostic dilemma may lead to confusion and uncertainty in starting ATT. The choice is left to the specialist, based on clinical acumen and the available epidemiological links, ATT is advocated. If available culture of the biopsy will help in confirmation of diagnosis. 

HIV-infected patients are generally undergoes annual tuberculosis skin test (or IGRA). What is your patient's condition? What is your patient's immunologic and virologic status? (HIV VL, CD4)

In the presence of a positive skin test along with clinical symptoms and positive imaging findings treatment could be start.

However, in any case, before treatment you will need further evaluation including gastric lavage for acid-fast smear, culture and PCR for Mycobacterium tuberculosis. Culture of sputum and BAL fluid for fungi, Mycobacterium tuberculosis and other mycobacteria. CMV PCR in BAL also may be helpfull in the case of early CMV pneumonitis.

Due to the CT finding non-tuberculosis mycobacterium also should be consider.

If this patient has normal CD4 counts, you should consider him a non-HIV patient with respect to tbc diagnosis. A true infiltrative process without mycobacteria is not quite usual presentation of tbc in a western country. Non-tb infection is still far more likely. What about skin testing and/or IGRA. Did you also perform TBB when performing bronchoscopy?

In haemoptysis do not forget to sample the expectorated blood as "sputum" for detection of AFB and culture. It comes from the site of potential infection and helps in the diagnosis.