The role and clinical significance of circulating tumor cells and of tumor DNA in the plasma have not yet been clarified. In the present study, we compared rates of detection of tumor-derived DNA in the buffy coat to those in plasma from tumor-bearing rats, and we attempted to correlate these rates with the progression of tumors. We injected DHD/K12-PROb cancer cells subcutaneously into BD-IX rats and divided the animals into six groups according to the time between the injection of tumor cells and euthanasia. After euthanasia, macroscopic metastases were assessed and samples of blood and lung were collected. We used mutant allele-specific amplification by PCR to detect tumor-derived DNA. We detected tumor DNA in lung samples from the first week after inoculation, in plasma from the third week and in the buffy coat from the fifth week. All animals analyzed on the 11th week had macro- or micrometastases in their lungs. Regardless of group, the rate of PCR-positive plasma samples was significantly higher than that of circulating tumor cells (P=0.005). In animals with metastases, this difference was also statistically significant (P=0.008). However, neither the detection of tumor DNA in the plasma nor the presence of circulating tumor cells was strongly correlated with the presence of metastases. Thus, cell-free tumor DNA was detected sooner and more frequently than circulating tumor cells and the dissemination of tumor DNA in the plasma seems to be much more common than detectable hematogenic tumor cells during the spread of colorectal cancer.