For the scheduled future revision of the TNM staging system for lung cancer, it is important that the present 1997 version be evaluated in a large population. In 2001, the Japanese Joint Committee of Lung Cancer Registry sent a questionnaire to 320 Japanese institutions regarding the prognosis and clinicopathological profiles of patients who underwent the resection for primary lung neoplasms in 1994. We compiled the data for 7408 patients from 303 institutions (94.7%). Among these, 6644 patients with non-small cell histology were studied in terms of prognosis. The 5-year survival rate of the entire group was 52.6%. The 5-year survival rates by clinical (c-) stage were as follows: 72.1% for IA (n = 2423), 49.9% for IB (n = 1542), 48.7% for IIA (n = 150), 40.6% for IIB (n = 746), 35.8% for IIIA (n = 1270), 28.0% for IIIB (n = 366) and 20.8% for IV (n = 147). The difference in prognosis between neighboring stages was significant except for between IB and IIA and between IIIB and IV. The 5-year survival rates by pathological (p-) stage were as follows: 79.5% for IA (n = 2009), 60.1% for IB (n = 1418), 59.9% for IIA (n = 232), 42.2% for IIB (n = 757), 29.8% for IIIA (n = 1250), 19.3% for IIIB (n = 719) and 20.0% for IV (n = 259). The difference in prognosis between neighboring stages was significant except for between IB and IIA and between IIIB and IV. The survival curves of stages IB and IIA were almost superimposed in both c- and p-settings. These findings indicated that the present stages IB and IIA should be merged into the same stage category. Otherwise, the present TNM staging system seemed to well characterize the stage-specific prognosis in non-small cell lung cancer. The future revision should focus on the subdivision of stages I and II.