Purpose: Advanced stage breast cancer patients may benefit from high dose chemotherapy, but only a minority will respond. A method to select non-responders as early as possible is essential for preventing unnecessary toxicity, possibly enabling a switch to an alternative treatment. Positron emission tomography (PET), a noninvasive molecular imaging modality, is able to detect functional changes well before anatomical changes are visible. The purpose of the present study was to develop a rat model of human breast cancer suitable for PET. This model, together with PET, would provide a means for investigating the efficacy of new anti-cancer drugs in an experimental setting.Methods: Human breast cancer cells MDA MB231 were injected subcutaneously into nude rats. Tumor take, tumor doubling time and growth inhibition after treatment with maximum tolerable doses of 5-fluorouracil, doxorubicin, cyclophosphamide and paclitaxel were established. As thymidine competes with 18 FLT and plasma thymidine levels are high in rodents, this could affect 18 FLT uptake. Therefore, use of thymidine phosphorylase to lower plasma thymidine levels was investigated. Finally, as an illustration of the potential use of the model, a pilot PET study was performed using 18 FDG and 18 FLT.Results: Tumor take rate was 68% when matrigel was coinjected. Tumor doubling time was 6 days. Treating animals with anti-cancer drugs resulted in tumor growth inhibition ranging from 7 to 68%, depending on the type of drug. Using thymidine phosphorylase, plasma concentrations of thymidine could be decreased by more than 80% and these reduced levels were stable for more than an hour, i.e., long enough for a PET study. Tumors could clearly be visualised using 18 FDG and 18 FLT PET.Conclusions: A new in vivo breast cancer model was successfully established in nude rats, allowing for quantitative PET studies of anti-cancer agents.