Introduction: We compared the accuracy of clinical nodal (cN) status N0-1 with that of pathological nodal (pN) status obtained from systematic mediastinal lymph node dissection (SMLD) in primary non-small cell lung cancer.
Methods: Data from 22 consecutive patients, who underwent lung cancer resection and SMLD of at least three mediastinal lymph node stations, from November 2001 to May 2003, were ana1ysed retrospectively. Only patients with cN0-1 status on computed tomography (CT) referred for surgery, were included in this study.
Results: Mean age of patients was 66.6 +/- 8.1 years with a male to female ratio of 17:5. Mean number of lymph node stations dissected was 5.8 +/- 1.8. 41 percent had squamous cell carcinoma, 45.5 percent had adenocarcinoma, and 4.5 percent each had large cell carcinoma, bronchioalveolar carcinoma or a lymphoepithelial carcinoma. pN2 metastases were found in 27.3 percent of patients. The sensitivity of cN0-1 was only 12.5 percent, with a specificity of 92.9 percent and an area under the receiver operating characteristics curve of 0.53. The positive and negative predictive values of cN0-1 status were 50 percent and 65 percent, respectively, with an accuracy of 59 percent. 41 percent of patients were understaged with 27.3 percent in pathological stage III. Curative resections were achieved in 59 percent of patients.
Conclusion: The sensitivity of cN0-1 status based on CT alone is extremely poor when compared with pN status from SMLD. Based on cN0-1 status alone without SMLD, 27.3 percent of patients in pN2 would have been understaged. We recommend that all patients with cN0-1 status should undergo SMLD of at least three appropriate mediastinal node stations, for more accurate staging.