Experience with Flexible Bronchoscopy in the Nonresolving Pulmonary Infiltrate

Flexible bronchoscopy (FFB) has been used widely in both immunocompromised patients, in whom opportunistic infection is the primary concern, and immunocompetent patients, in whom infection and malignancy are issues. The operating characteristics of FFB in these diverse settings have not been fully explored, prompting us to review our community's experience with the procedure. We identified patients under-going FFB for a nonresolving infiltrate over a 2-year period at four Rochester, New York hospitals. We classified the patients into 3 groups: Group I, immunocompromised other than acquired immunodeficiency syndrome (AIDS): Group II, immunocompetent with symptoms of pneumonia for at least 10 days with prior antibiotic therapy of at least a week; and Group III, immunocompetent patients suspected of having pulmonary malignancy. We determined the operating characteristics of FFB by pursuing follow-up until a diagnosis was established or for a minimum of 2 years by review of records, and by contacting the patient or attending physician as necessary. We were successful in follow-up on 159 patients (31 Group I, 64 Group II. and 64 Group III). FFB was most likely to yield a specific diagnosis in Group I patients (61%), compared to 38% and 31% for Groups II and III, respectively. The infiltrate in the immunocompetent patient was most likely to he infection when features of pneumonia were present (Group II), and neoplastic when not (Group III). We conclude the specificity and positive predictive values of FFB were excellent for all 3 groups. The sensitivity of the procedure was best for Group II patients (92%). All Group III patients with a false negative result were subsequently determined to have carcinoma. The predictive value of a negative result and hence the diagnostic value of FFB was best for immunocompetent patients, especially when features of pneumonia were present (95%). Despite a higher diagnostic yield, the value of a negative result was least for the immunoincompetent group (50%). We conclude that a patient's clinical features at presentation define the utility of FFB for those with a nonresolving infiltrate.