Amplification of cytokeratin 19 (CK19) transcripts by reverse transcriptase-polymerase chain reaction (RT-PCR) has been shown to be a highly sensitive assay for the detection of bone marrow micrometastases (BMM) of breast cancer, but recent studies have demonstrated the occurrence of false-positive results due to low-level, illegitimately transcribed CK19 in normal bone marrow tissue. One approach to solve this problem is to develop a quantitative CK19 RT-PCR assay and to introduce a cut-off value which can distinguish between illegitimate expression and cancer-specific expression levels. In the present paper, we describe a quantitative CK19 RT-PCR assay using a real-time automated PCR system. The number of CK19 transcripts was normalized to that of GAPDH transcripts as an internal control for quality and quantity of cDNA. The cut-off value for the ratio of CK19 to GAPDH transcripts was set at 10(-4) since the ratio never exceeded this value in the control bone marrow samples (n = 12). In total, 117 bone marrow aspirates from stage I - III patients with invasive breast cancers were subjected to CK19 RT-PCR assay and immunocytological examination. Forty (34.2%) were found to be BMM-positive by CK19 RT-PCR assay whereas only three (2.6%) were found to be BMM-positive by immunocytology. Multivariate analysis has shown that occult BMM detected by CK19 RT-PCR is a significant risk factor for relapse, being independent of axillary lymph node metastases.