Methodological Analysis of Immunocytochemical Screening for Disseminated Epithelial Tumor Cells in Bone Marrow

Abstract

The emerging clinical relevance of bone marrow micrometastasis has prompted several investigations, using a variety of immunocytochemical approaches. The present study was designed to evaluate some of the variables affecting the immunocytochemical detection of individual epithelial tumor cells in bone marrow. Using an alkaline phosphatase-antialkaline phosphatase staining technique, we evaluated bone marrow aspirates from 358 patients with primary carcinomas of the breast (n = 150), lung (n = 66), prostate (n = 42), or colorectum (n = 100). Individual tumor cells in cytological preparations were detected with monoclonal antibody (MAb) CK2 to the epithelial cytokeratin component 18 (CK18), which has been validated in extensive clinical studies. In addition, the utility of the broad-spectrum MAb A45-B/B3 was explored in this study. The high specificity of MAbs CK2 and A45-B/B3 was supported by analysis of bone marrow from 75 noncarcinoma control patients and by double-marker analysis with MAbs to mesenchymal marker proteins (CD45 and vimentin). In contrast, MAbs E29 and HMFG1, directed to mucin-like epithelial membrane proteins, cross-reacted with hematopoietic cells in 26.7-42.7% of all samples tested. The majority of the 154 positive samples (43.0%) from cancer patients displayed less than 10 CK18-positive cells per 8 x 10⁵ marrow cells analyzed. The detection rate, however, was affected by blood contamination of the aspirate, the number of aspirates analyzed, and the number of marrow cells screened per aspiration site. Comparative immunostaining of bone marrow specimens with MAbs CK2 and A45-B/B3 indicated that downregulation of CK18 in micrometastatic carcinoma cells occurs in about 50% of the 172 samples analyzed, regardless of the primary tumor origin. In conclusion, the result of the immunocytochemical screening for single carcinoma cells in bone marrow largely depends on the method applied, which underscores the urgent need for a standardized protocol. At present, we recommend screening of at least 2 x 10⁶ nucleated cells aspirated from two sites of the iliac crest, using a broad-spectrum cytokeratin MAb with proven specificity for epithelial cells.