Immunocytological detection of bone marrow micrometastasis in operable non-small cell lung cancer.

Present diagnostic techniques do not allow the detection of early metastatic spread of tumor cells, although this spread largely determines the clinical course of patients with small primary cancers. By use of monoclonal antibody CK2 to the epithelial cytokeratin component number 18 (CK18), individual disseminated carcinoma cells present in bone marrow of cancer patients can now be identified (G. Schlimok, I. Funke, B. Holzmann, G. Göttlinger, G. Schmidt, H. Häuser, S. Swierkot, H. H. Warnecke, B. Schneider, H. Koprowski, and G. Riethmüller, Proc. Natl. Acad. Sci. USA, 84: 8672-8676, 1987; F. Lindemann, G. Schlimok, P. Dirschedl, J. Witte, and G. Riethmüller, Lancet, 340: 685-689, 1992). In the present study, we applied this approach to patients with operable non-small cell lung cancer. CK18 was expressed on 84 of 88 (95.5%) primary adenocarcinomas and squamous cell carcinomas. Irrespective of primary tumor histology, single aspirates of iliac bone marrow from 18 of 82 (21.9%) lung cancer patients exhibited between 1 and 531 CK18+ cells/4 x 10(5) nucleated marrow cells. The specificity of our assay is underlined by the small rate of "false-positive" cells being observed in only 2 of 117 (1.7%) marrow samples from control patients with no evidence for an epithelial malignancy at the time of aspiration. Comparison with established risk factors demonstrated positive correlations (P < 0.05) between the size and histological grade of the primary carcinoma and cytokeratin positivity in iliac bone marrow. In contrast, the association with the metastatic involvement of regional lymph nodes was only weak (P = 0.09). Following a median observation period of 13 months, patients who displayed cytokeratin-positive cells in iliac bone marrow at the time of primary surgery relapsed more frequently as compared to patients with a negative marrow finding (66.7 versus 36.6%; P < 0.05). This difference was even more pronounced by comparing the rates of manifest skeleton metastasis observed in both groups (26.7 versus 2.4%; P < 0.005). Finally, colabeling of CK18+ cells in marrow with monoclonal antibodies to proliferation-associated markers, such as the nucleolar antigen p120 or the tyrosine kinase receptor erbB2, exemplified the oncogenic capacity of CK18+ micrometastatic cells. In conclusion, CK18+ cells present in the bone marrow of patients with apparently operable non-small cell lung cancer exhibit the potential to form solid metastases. Therefore, the approach presented here may be used to determine the risk of early relapse in operable non-small cell lung cancer with potential consequences for adjuvant therapy.