BACKGROUNDThe accuracy of clinical TNM staging correlates with appropriate treatment in patients with lung cancer. The aim of this study is to evaluate the agreement between clinical and surgical-pathological staging in patients with lung cancer and analyze its cause in detail.

METHODSOne hundred and fifty patients with lung cancer treated surgically from 2000 were enrolled randomly. Clinical and surgical-pathological staging of them were made respectively according to the International System for Staging Lung Cancer newly revised by UICC. Then concordance was determined between the two staging results with Kappa value, and difference in coincident rate was analyzed among subgroups of T staging.

RESULTSFor T staging, the agreement was excellent (Kappa value=0.729), however, the coincident rate of T3 or T4 was significantly lower than that in T1 or T2 group (P < 0.01). The agreement of N staging was poor (Kappa value=0.108), followed by a disappointed conformity of TNM staging (Kappa value=0.287).

CONCLUSIONSClinical T staging based on CT can indicate the location and size of primary tumor precisely. But the borderline may be difficult to estimate when tumor site is near chest wall or mediastinum, so some patients with clinical T4 still have chances to receive complete resection. The conformity of N staging is rather poor. The key point to improve the accuracy of clinical TNM staging should be to seek more reliable techniques for evaluating N status.