We previously reported in a retrospective study that CEA-based RT-PCR was useful for predicting the prognosis of patients with node-negative colorectal cancer. RT-PCR is well established for laboratory use, but many issues remain to be resolved prior to its clinical application. In addition to the false positive rate of RT-PCR, we addressed several issues, including the timing of lymph node sampling, stability of RNA after surgery, and reproducibility of results. After appropriate modification, including development of a tissue sampling kit, a multi-institutional clinical study was commenced prospectively from November 2001, and 100 patients were enrolled for examination of micrometastasis. RNA was stable in lymph nodes for up to 3 h after surgical resection. This range of sampling time was acceptable to the surgeons. RNA was well preserved in RNA later at -20 degrees C for 3 weeks. Dilutions of MKN45 and LoVo cells served as positive controls for conventional PCR since these controls were found to be highly stable and ensured reproducibility. Moreover, simultaneous use of quantitative PCR (Light Cycler) ensured double confirmation of the results. Our clinical study showed that the quality of RNA was excellent or good in most samples (98 of 100; 98%). Twenty-four of 98 (24.5%) cases were judged to be micrometastasis-positive. In conclusion, the current translational research study established a clinically feasible RT-PCR system for micrometastasis. Our system could potentially be useful as a clinical tool.