The aim of this study was to evaluate the prognosis for patients with lung cancer detected by helical CT but not by CXR. One hundred and thirty-seven asymptomatic patients with lung cancer diagnosed by annual mass screening of the chest were enrolled over a 7-year period. Five-year survival rates in patients with lung cancer detected only by helical CT (n = 19: CT-only detection group) and in patients with lung cancer visible by both CXR and helical CT (n = 118: control group) were evaluated, and clinical variables were examined as possible predictors of survival time using the Cox proportional-hazards model. There was a significant difference between the 5-year survival rates in the CT-only detection group and in the control group (80% vs. 39%, log rank: P = 0.0171). The risk of death decreased 77% in CT-only detectable lung cancer (hazard ratio: 0.219, 95% confidence interval: 0.057-0.845, P = 0.0275). Lung cancer could not be seen by CXR because nodules were small or faint (n = 11) or overlapping a shadow of thoracic components (n = 8). The percentage of subsolid nodules (classified as either part-solid or non-solid nodules) was higher in the subgroup with small or faint nodules (82% vs. 25%, P = 0.0423). Helical CT has the ability to detect early lung cancer before the small or faint nodules increase to a size visible on CXR, and patients with lung cancer detected only by helical CT have a better prognosis.