Does the use of immunohistochemistry to identify micrometastases provide useful information in the staging of node-negative non-small cell lung carcinomas?

Immunohistochemical studies using epithelial markers have recently been published which identified micrometastases in lymph nodes that had not been found on routine pathological assessment, therefore increasing the accuracy of staging of non-small cell lung cancers. The presence of these micrometastases was associated with reduced survival. We have therefore performed a retrospective immunohistochemical study on all the lymphoid tissue from five lymph node stations (2 hilar, 3 mediastinal) from 49 patients with T1-2, NO disease. Before immunohistochemistry was undertaken, all slides were reviewed, with the lymph nodes confirmed as negative. In total, 1447 lymph node slices (average 30 per case, 5.9 per lymph node station) were examined, these figures reflecting sectioning of lymph nodes at approximately 3 mm intervals before processing. MNF116, a broad spectrum anti-keratin antibody was then used to look for occult metastases, with adjacent serial sections being examined to ensure that any positively staining cells were detected solely by immunohistochemistry and not through deeper sectioning. In five cases, lymph nodes contained positively staining cells. Two cases proved to be false positives, further immunohistochemistry identifying the cells as benign mesothelial inclusions. In the remaining three cases, positive staining correlated with tumour cells in the adjacent serial sections. Follow-up on 46 of 49 patients revealed recurrence in 27% (actual survival 68%); however all three cases containing tumour cells on immunohistochemistry were free from recurrence. These results suggest that the use of immunohistochemistry adds little useful information above that of thorough routine examination of lymph nodes. They also document that benign mesothelial inclusions within lymph nodes are more frequent than previously reported. (C) 1997 Elsevier Science Ireland Ltd.