of Work: This project is a test of the efficacy of a new diagnostic method for imaging the airways known as virtual bronchoscopy. Virtual bronchoscopy is performed by acquiring thin-section computed tomography (CT) images of the chest. These images are used to generate a three-dimensional model of the tracheal and bronchial walls on a graphics workstation. The model can be manipulated to allow the viewer to """"""""fly through"""""""" the tracheobronchial tree providing views similar to those obtained during bronchoscopy. The technique produces a display of the human bronchial system in a readily understood format. Moreover, it allows investigation of post-stenotic portions of the bronchial tree that are beyond the reach of fiber-optic bronchoscopy. The patients in this protocol have inflammatory, infectious, or neoplastic pulmonary processes and would have had chest CT for clinical reasons. Over the past 2 years, we have developed software techniques for navigating inside virtual bronchoscopy studies and have developed image processing techniques to improve the efficiency and accuracy of these studies. We are also in the process of accruing patients who have had both fiber optic bronchoscopy and virtual bronchoscopy in order to compare the two methods. We have performed a descriptive analysis of cavitary lung lesions. Finally, we have developed automated techniques to assist physicians in locating and characterizing endobronchial lesions. We anticipate that virtual bronchoscopy will be diagnostically efficacious for disorders that produce a morphologic alteration of bronchial anatomy.
