The international tumor-node-metastasis (TNM) staging system is the "international language" in cancer diagnosis and treatment. Six revisions of the TNM staging system for lung cancer have been repeated over the past 35 years after the beginning of UICC-TNM classification in 1968. The 1997 revision for lung cancer has undergone an extensive correction for many deficiencies of the old staging system. As a result, the new staging system appears to be a great improvement over previous editions. There are, however, still some controversies and proposals for revising, even when the new staging system is applied in daily diagnoses and treatment for lung cancer. In the present paper, these problems are presented and discussed. Main subjects for discussions are as follows: (1). Since the 2nd revision, T1 and T2 lesions were divided at the border of a 3 cm tumor size. Is 3 cm diameter an appropriate cut-off point for dividing T1 and T2 lesions? (2). Is it valid to subdivide T1 and T2 lesions into each A and B? (3). Is it appropriate to down-stage all of T3N0M0 to stage IIB, because there exists heterogeneity of T3? (4). Definitions of T4 lesion. (5). Controversies in three kinds of lymph node maps. Especially, where there is a boundary between N1 and N2 station in each map? (6) How to classify separate tumor nodules (STN) in the same lobe, and in the non-primary lobe. (7) Controversy exists concerning the validity of present stage grouping, because there are no significant difference of survivals between IB and IIA, IIA and IIB in most reports and also between T3N0M0 and T3N1M0 in some reports.