OBJECTIVE

The 1997 version of the TNM staging system for lung cancer has several prognostic problems. Among these, the overlapping survival of stages IB and IIA is the most serious. We performed this retrospective study to test a revised TNM staging system for lung cancer.

METHODS

We revised the T1 descriptor definition and stage grouping for testing as follows. According to the greatest tumor diameter, T1 tumors were divided into T1a tumors (< or =2.0 cm) and T1b tumors (2.1-3.0 cm). With these descriptors, new IA, IB, and IIA stages were defined as T1a N0 M0, T1b N0 M0, and T2 N0 M0 + T1 N1 M0, respectively. For 6644 patients with histologically non-small cell lung cancers resected in 1994 and reported in the Japanese Lung Cancer Registry Study, the survivals and prognostic difference between neighboring stages were studied.

RESULTS

The 5-year survival of the entire population was 52.6%. In the clinical setting, the 5-year survivals of the new IA, new IB, new IIA, IIB, IIIA, IIIB, and IV stages were 77.5%, 69.3%, 49.8%, 40.6%, 35.8%, 28.0%, and 20.8%, respectively. In the pathologic setting, they were 83.7%, 76.0%, 60.0%, 42.2%, 29.8%, 19.3%, and 20.0%, respectively. For both clinical and pathologic settings, differences between all neighboring stages were statistically significant, except for that between IIIB and IV.

CONCLUSION

Subcategorization of T1 and minor changes in stage grouping results in a system with significant differences in prognosis between neighboring stages. The unification of stages IB and IIA, especially, improves the discriminatory power of the staging system.