The detection of micrometastasis holds great promise for earlier staging of patients with malignant diseases and may ultimately guide therapeutic decisions. So far, reverse-transcriptase polymerase chain reaction (RT-PCR) amplification of genes expressed by the tumor in a tissue-specific manner is the method with the highest diagnostic sensitivity. It is well-established that the identification of single tumor cells is feasible in tissues and bodily fluids in both experimental and clinical samples. However, at present it is difficult to assign clinical significance to results obtained from such tests, primarily because their diagnostic specificity is disputed, both conceptionally and methodologically. For example, amplification of candidate mRNA targets is detectable in non-cancer patients using conditions that generally fail to generate such signals from healthy individuals. We have established that transcription of the tissue-specific genes can be affected by different means. Specifically, some target mRNA species are detectable in peripheral blood nuclear cells as low abundance constitutive-like expression, whereas others are induced through in vitro tissue culturing. In addition, mRNA expression may be distinctly upregulated by different cytokines or growth factors in vivo. Also, background transcription of target mRNAs can occur in different lineages of peripheral blood cells. Finally, expression may be substantially different in tissues such as peripheral blood, bone marrow, or lymph nodes. As a consequence, cancer patients in unrelated clinical situations may present with different levels of background expression, making the diagnostic specificity of test results difficult to assess. To add to this complexity, an increasing body of literature is being generated using various targets for a multitude of malignant diseases. There is a great variety of methods for sampling, specimen processing, nucleic acids recovery, test conditions, and readout formats, making it impossible to compare data. In summary, modalities of quantitative RT-PCR methods and standardization issues should be discussed to address these questions.