Given the variability in rate of radiographic resolution, it remains controversial to decide when to initiate an invasive diagnostic work-up for nonresolving or slowly resolving pulmonary infiltrates. In immunocompetent patients who present with classical features of CAP (i.e., fever, chills, productive cough, new pulmonary infiltrate), clinical response to therapy is the most important determinant for further diagnostic studies. Within the first few days, persistence or even progression of infiltrates on chest radiographs is not unusual. Defervescence, diminished symptoms, and resolution of leukocytosis strongly support a response to antibiotic therapy, even when chest radiographic abnormalities persist. In this context, observation alone is reasonable, and invasive procedures can be deferred. Serial radiographs and clinical examinations dictate subsequent evaluation. In contrast, when clinical improvement has not occurred and chest radiographs are unchanged or worse, a more aggressive approach is warranted. In this setting, we advise fiberoptic bronchoscopy with BAL and appropriate cultures for bacteria, legionella, fungi, and mycobacteria. When endobronchial anatomy is normal and there is no purulence to suggest infection, TBBs should be done to exclude noninfectious causes (discussed earlier) or infections attributable to mycobacteria or fungi. An aggressive approach is also warranted in patients who are clinically stable or improving when the rate of radiographic resolution is delayed. As discussed earlier, what constitutes excessive delay is controversial, and depends upon the acuity of illness, specific pathogen, extent of involvement (i.e., lobar versus multilobar), comorbidities, and diverse host factors. Stable infiltrates even 2 to 4 weeks after institution of antibiotic therapy does not mandate intervention provided patients are improving clinically. Invasive techniques can also be deferred when unequivocal, albeit incomplete, radiographic resolution can be demonstrated. Lack of at least partial radiographic resolution by 6 weeks, even in asymptomatic patients, however, deserves consideration of alternative causes (e.g., endobronchial obstructing lesions, or noninfectious causes). Fiberoptic bronchoscopy with BAL and TBBs has minimal morbidity and is the preferred initial invasive procedure for detecting endobronchial lesions or substantiating noninfectious causes. The yield of bronchoscopy depends on demographics, radiographic features, and pre-test likelihood. In the absence of specific risk factors, the incidence of obstructing lesions (e.g., bronchogenic carcinomas, bronchial adenomas, obstructive foreign body) is low. Bronchogenic carcinoma is rare in nonsmoking, young (< 50 years) patients but is a legitimate consideration in older patients with a history of tobacco abuse. Non-neoplastic causes (e.g., pulmonary vasculitis, hypersensitivity pneumonia, etc.) should be considered when specific features are present (e.g., hematuria, appropriate epidemiologic exposures). Ancillary serologic tests or biopsies of extrapulmonary sites are invaluable in some cases. In rare instances, surgical (open or VATS) biopsy is necessary to diagnose refractory or non-resolving "pneumonias."