Abstract
Introduction: Virtual bronchoscopy (VB) is a new, noninvasive diagnostic technique that allows visualization of trachea and bronchi. Virtual images are created based on scans from helical multidetector computed tomography (CT) scans using a special protocol. The obtained reconstructed virtual image of the bronchial tree reflects pictures seen under conventional fibre-optic bronchoscopy (FOB). The aim of the study was to compare VB images of the bronchial tree with findings from FOB and to evaluate the diagnostic value of VB in the examination of central lung tumours. Material and methods: The studied group consisted of 40 patients with suspected central lung tumours, detected first on chest X-ray. Each patient underwent routine CT and FOB, followed by VB. Physicians performing FOB were blinded to VB results and vice versa. The presence of tumour, bronchial stenosis, and widening of the carina were the analysed features. Lung cancer was confirmed by histopathological examination in all patients, including 32 cases of non-small cell lung cancer (80%), 2 cases of small cell lung cancer (5%), 5 cases of squamous cell carcinoma (12.5%), and one case of carcinoid (2.5%). Results: Virtual bronchoscopy for bronchial tumour detection had a sensitivity of 79.5% and specificity of 95.5%. When bronchial stenosis was assessed, sensitivity was 58.6% and specificity was 98.1%, whereas detection of widening of carina had a sensitivity of 60.7% and specificity of 97.7%. Conclusions: The results indicate that virtual bronchoscopy is a highly sensitive and specific diagnostic method, of high clinical importance in the evaluation of lung tumours with central location.