Transtracheal selective bronchial brushing for pulmonary infiltrates in patients with cancer.

Patients with cancer frequently develop pneumonitis for which no cause is documented ante mortem. Noninvasive diagnostic techniques, such as sputum induction, are generally inadequate, especially in myelosuppressed patients. To avoid pulmonary contamination with organisms colonizing the oronasopharynx and to obtain uncontaminated speciemens, 38 patients underwent bronchial brushing utilizing a transtracheal approach after sputum induction and transtracheal aspiration failed to establish the etiology. Patients with thrombocytopenia were brushed after platelet transfusion. Eleven patients were not clinically considered to be infected; seven proved to have pulmonary metastases, of which one case was diagnosed by this technique; and four patients in whom no diagnosis was obtained by brushing subsequently proved to have interstitial fibrosis (three cases) or a collapsed lobe (one case). Twenty-seven patients were clinically presumed to be infected. Ultimately, 17 of these 27 patients were proven to have pulmonary infection, and 14 of these 17 were etiologically documented by brushing. In ten of the 27 patients presumed to be infected, no etiology could be established by any method. Seven of these ten patients were receiving broad-spectrum antibiotic therapy at the time. Significant but nonfatal complications, including hemoptysis, pneumothorax, and cervical cellulitis, occurred in seven patients; however, this procedure is a relatively safe and useful method to include in the orderly evaluation of myelosuppressed cancer patients with suspected pulmonary infections.