The aim of this study was analysis of incidence results in a prospective one-arm feasibility study of lung cancer screening with low-radiation-dose spiral computed tomography in heavy smokers. Eight hundred seventeen smokers (> or =40 years, > or =20 pack years of smoking history) underwent baseline low-dose CT. Biopsy was recommended in nodules >10 mm with CT morphology suggesting malignancy. In all other lesions follow-up with low-dose CT was recommended. Annual repeat CT was offered to all study participants. Six hundred sixty-eight (81.8%) of the 817 subjects underwent annual repeat CT with a total of 1735 follow-up years. Follow-up of non-calcified nodules present at baseline CT demonstrated growth in 11 of 792 subjects. Biopsy was performed in 8 of 11 growing nodules 7 of which represented lung cancer. Of 174 new nodules, 3 represented lung cancer. The 10 screen-detected lung cancers were all non-small cell cancer (6 stage IA, 1 stage IB, 1 stage IIIA, 2 stage IV). Five symptom-diagnosed cancers (2 small cell lung cancer: 1 limited disease, 1 extensive disease, 3 central/endobronchial non-small cell lung cancer, 2 stage IIIA, 1 stage IIIB) were diagnosed because of symptoms in the 12-month interval between two annual CT scans. Incidence of lung cancer was lower than prevalence, screen-detected cancers were smaller, and stage I was found in 70% (7 of 10) of screen-detected tumors. Only 27% (4 of 15) of invasive procedures was performed for benign lesions; however, 33% (5 of 15) of all cancers diagnosed in the population were symptom-diagnosed cancers (3 central NSCLC, all stage III, 2 SCLC) demonstrating the limitations of CT screening.