Endobronchial tumours in a campaign for early detection of bronchial cancer: Computed tomography versus endoscopy

Abstract Objective To study endobronchial cancers occurring in a population at high risk of bronchial cancer (history of surgically treated bronchial or ENT cancer in complete remission, and symptoms due to smoking) detected by annual volume CT scans and biannual fibroscopy. Material and methods Two hundred and sixty-six patients were included in this single centre prospective study; 27 bronchopulmonary cancers were detected. Ten endobronchial cancers (37%) were identified by fibroscopy (nine invasive cancers and one carcinoma in situ) in 10 patients (nine men) (51–78 years old) nine of whom were smokers (dark tobacco: seven). The screening CTs were reappraised by two radiologists. Results Three cancers out of 10 were detected by CT during the initial reading. The sensitivity of the reappraised CT was 80% with seven false positives. In five cases, the mean period between the first CT scan where the lesion was visible retrospectively, but not described, and the diagnostic fibroscopy was 463 days (213–808 days); two cancers were not visible in the CT scan. Seven curative treatments were undertaken. Conclusion In this population, the sensitivity of the initial reading of the CT scan for detecting endobronchial tumours was 30%, while 80% of the tumours were visible retrospectively, underlining the importance of careful analysis of the proximal bronchial tree.