Comparison of computed tomography and systematic lymph node dissection in determining TNM and stage in non-small cell lung cancer.

Objective: To compare computed tomography (CT)-based clinical TNM and staging to surgical-pathological staging with systematic lymph node dissection in primary non-small cell lung cancer.

Methods: The study included 49 non-small cell lung cancer patients that underwent lung resection and systematic lymph node dissection between 1997 and 2001. Preoperative clinical and CT findings were compared with surgical-pathological findings. Lymph nodes with a shortest diameter of over 1 cm on CT were considered abnormal, but did not contraindicate surgery. Patients with CT indicating an invasive T4 tumor, pleural carcinosis, or bulky N2 disease were excluded.

Results: Sixty-five percent (32/49) had epidermoid carcinoma, and 25% (12/49) had adenocarcinoma. N2 metastases were found in 12% (6/49). The clinical T category was correct in 71% (35/49), and the N category in 55% (27/49). The sensitivity for detecting N2 disease was 67% (4/6), and the specificity was 81% (35/43). The positive predictive value for N2 disease was 33% (4/12), and the negative predictive value was 95% (35/37). Node-by-node agreement on N2 metastatic location was 17% (1/6). Skip N2 metastases without any N1 involvement were found in 4% (2/49), or 33% (2/6) of all N2 cases. The clinical stage was correct in 45% (22/49), and complete TNM agreement was 37% (18/49).

Conclusions: The clinical TNM and staging based on CT are inaccurate. The sensitivity for detecting N2 disease is poor, especially on node-by node basis. Preoperative exclusion of N2 metastases is quite reliable, but a positive finding should always be verified. Systematic mediastinal lymph node dissection is necessary to detect N2 metastases inaccessible to cervical mediastinoscopy, and skip N2 metastases without N1 involvement.