Introduction: In early stage non-small cell lung cancer (NSCLC), presence of lymphatic micrometastases and isolated tumor cells, primarily detected by immunohistochemistry, is suggested to be a prognostic factor. However, there is no consensus whether immunohistochemistry should be used routinely in lymph node assessment. The goal of our study was to determine whether recurrent disease is associated with the presence of lymphatic micrometastases and/or isolated tumor cells, at the time of the lung resection. Methods: We retrospectively analyzed the prevalence of lymphatic micrometastases and/or isolated tumor cells in two groups of patients, who underwent a curative resection for early stage NSCLC. Group I had a follow-up of 5 years without recurrent disease. Group II consisted of a matched group of patients with recurrent disease. Patients were originally classified as having negative mediastinal lymph nodes. All lymph nodes obtained by mediastinoscopy and thoracotomy were re-examined by serial sectioning and immunohistochemistry. Results: Micrometastases and/or isolated tumor cells were found in one of 16 patients in group I, which was significantly different from six of 16 patients in group II. (Fisher exact test, 4.6; p, 0.04; risk ratio, 2.4). Serial sectioning and immunohistochemistry did not change N-stage for the single patient in group I, in contrast to all six patients in group II. Conclusion: Presence of lymphatic micrometastases and/or isolated tumor cells is associated with distant recurrence in patients with early stage NSCLC. We recommend the routine use of serial sectioning and immunohistochemistry in lymph node assessment to improve the accuracy of staging.