The aim of our study was to detect micrometastatic breast cancer by epithelial glycoprotein-2 (EGP-2) and cytokeratin 19 (CK19), using immunostaining and real time quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR). Fifty-eight breast cancer patients, 52 primary tumors, 75 sentinel nodes (SN) and 149 peripheral blood (PB) samples (from before, during and 4 days after operation) were examined. Immunostaining was performed with antibodies directed against EGP-2 and CK19. Detection limits were one Michigan Cancer Foundation-7 (MCF-7) breast cancer cell line cell/2.10(6) leukocytes (immunostaining) and one MCF-7 cell/10(6) leukocytes qRT-PCR. Control noncancer lymph nodes (n = 10) showed nonspecific CK19 staining, but were qRT-PCR negative; control healthy volunteer PB (n = 11) was always negative. Primary tumor samples, all positive with immunostaining, showed a wide variation of EGP-2 (>10(4) fold) and CK19 mRNA expression (>10(3) fold). SN (n = 19) from 16 patients were tumor-positive with routine haematoxylin-eosin (H&E) and/or immunostaining. SN tumor presence was positively correlated to qRT-PCR expression, but 3 tumor-positive SN were false negative with qRT-PCR. Three SN were qRT-PCR positive, while tumor negative with H&E and/or immunostaining. No immunostaining positive PB was observed, but 19 patients (33%) had one or more qRT-PCR positive PB samples. We concluded that primary tumors have varying expressions of EGP-2 and CK19 mRNA. Both markers can be used in qRT-PCR to obtain adequate sensitivity for single tumor cell detection. In SN, immunostaining appears more sensitive/specific than H&E or qRT-PCR for tumor detection. No immunostaining positivity was found in PB, while 33% of patients had qRT-PCR positive PB. The clinical value of these findings will have to be clarified.
Copyright 2003 Wiley-Liss, Inc.