Immunodeficient animals, principally nude mice, when used in appropriately designed studies have been shown to be useful for the experimental analysis of human breast cancer metastasis. As with many other human tumors, the implantation of breast cancer cells into an anatomically appropriate tissue (the mammary fatpad) results in increased tumor take and incidence of metastasis for certain cell lines compared with subcutaneous injection. Testing a number of widely available human breast cancer cell lines identified the MDA-MB-435 cell line as the most metastatic, producing lung and lymph node metastases in a high proportion of nude and severe combined immunodeficient (SCID) mice after injection in the mammary fatpad. Mixing human breast cancer cells with normal fibroblasts or with Matrigel also increases the tumor incidence and growth rates in nude mice. Different routes of injection can be used to assess the ability of human breast cancer cells to form metastatic lesions in the lungs (i.v. injection), the liver (injection in the spleen), the brain (direct or intracarotid artery injection) and the bone marrow and bone (injection into the left ventricle of the heart). These different approaches demonstrate the potential of experimental studies of human breast cancer growth and metastasis using immunodeficient mice; this model is valuable for experiments that test the role of metastasis-associated genes and the efficacy of novel forms of therapy.