Abstract

Computed tomography is widely used in the investigation of patients in whom bronchiectasis is suspected, despite considerable variation in its reported sensitivity and specificity. The findings with 3 mm high resolution computed tomography were compared at segmental level with bronchography by two radiologists independently in 27 patients (aged 20-67 years) undergoing investigation of chronic sputum production. Fifteen patients were found to have bronchiectasis by both investigations. Five were identified by computed tomography alone, including two in whom disease was revealed in segments underfilled at bronchography. The sensitivity of computed tomography compared with bronchography in the diagnosis of bronchiectasis at segmental level was 84% and the specificity 82%. The predictive value of computed tomography in the diagnosis of bronchiectasis was 38% overall, but increased to 75% when only those segmental bronchi moderately or severely dilated on the computed tomography scan were considered. There was no relation between the degree of bronchial wall thickening on the computed tomogram and the diagnosis of bronchiectasis by bronchography. Bronchography may be avoided in patients being considered for surgical resection of their bronchiectasis in whom computed tomography shows diffuse disease.