Multidetector CT virtual bronchoscopy to grade tracheobronchial stenosis.

OBJECTIVE: The purpose of this study was to compare the efficacy of noninvasive multidetector CT (virtual bronchoscopic images, axial CT slices, coronal reformatted images, and sagittal reformatted images) in depicting and allowing accurate grading of tracheobronchial stenosis with that of flexible bronchoscopy.
MATERIALS AND METHODS: Multidetector CT and flexible bronchoscopy were used to examine 200 bronchial sections obtained from 20 patients (15 patients with bronchial carcinoma and five without central airways disease). Multidetector CT was performed using the following parameters: collimation, 4 x 2 mm, pitch, 1.375; and reconstruction intervals, 2 mm. Postprocessing was performed using surface rendering and multiplanar reformatted images. CT images were independently interpreted by two radiologists. The tracheobronchial stenoses revealed on flexible bronchoscopy were graded by a pulmonologist.
RESULTS: Virtual bronchoscopic findings, axial CT scans, and multiplanar reformatted images were highly accurate (98% accuracy for virtual bronchoscopic images, 96% for axial slices and coronal reformatted images, and 96.5% for sagittal reformatted images) in revealing tracheobronchial stenosis. In allowing accurate grading of tracheobronchial stenosis, images from virtual bronchoscopy correlated closely (r = 0.91) with those of flexible bronchoscopy. Because use of virtual bronchoscopic images reduced the overestimation of stenosis, these images allowed better assessment of stenosis than did axial CT slices (r = 0.84) or multiplanar reformatted images (r = 0.84) alone.
CONCLUSION: Multidetector CT virtual bronchoscopy is a reliable noninvasive method that allows accurate grading of tracheobronchial stenosis. However, it should be combined with the interpretation of axial CT images and multiplanar reformatted images for evaluation of surrounding structures and optimal spatial orientation.