The introduction of nude mice to cancer research in 1969 by Rygaard allowed successful transplantation of human cancers. Initial models using nude mice involved subcutaneous transplantation of human tumors which did not metastasize. The development of orthotopic models in the late 1970s and early 1980s by Heppner, Sordat, and others demonstrated that spontaneous metastasis of human tumors could occur in nude mice. The work of Fidler in the late 1980s and 1990s put orthotopic models on a solid scientific basis and demonstrated their great potential for understanding metastasis. At the beginning of the 1990s, our laboratory introduced surgical orthotopic implantation (SOI) models, whereby tumor fragments are implanted orthotopically by surgical methods and enabled patient-like orthotopic nude mouse models to be developed for every major cancer type. With the SOI model, we demonstrated in 1995 that organ colonization is the governing step of metastasis. The discovery of green fluorescent protein (GFP) by Shimomura in 1962 and its cloning by Prasher, 30 years later, provided the opportunity for our laboratory to label metastatic cancer cells in vivo with GFP in 1997. With the use of GFP and other colored fluorescent proteins such as red fluorescent protein (RFP) discussed by Labas, metastasis could be visualized in living mice in orthotopic models, at cellular and subcellular resolution, in real time. The use of GFP- and RFP-labeled cancer cells enabled our laboratory, in collaboration with Glinsky in 2003, to observe that orthotopic tumor models produced highly-metastatic viable circulating tumor cells (CTC) and subcutaneous tumor models did not. These results may explain why orthotopic tumor models metastasize and subcutaneous tumor models do not. The introduction of GFP, RFP, and cyan fluorescent protein (CFP) transgenic nude mice by our laboratory, beginning in 2004, enabled color-coded imaging of cancer cell/stromal cell interaction during metastasis in orthotopic models. With these models, our laboratory demonstrated that tumor-associated stromal cells are necessary for metastasis in 2006 and in collaboration with Verma in 2011, demonstrated that cancer cells can transdifferentiate into vascular endothelial cells. The clinical potential of these models for metastatic cancer was demonstrated by our laboratory in 2009 with the use of a telomerase-dependent adenovirus to label metastatic cancer in situ which enabled fluorescence-guided resection of the metastatic nodules. Powerful technologies of orthotopic transplantation and fluorescent-protein imaging have brought metastasis research to a level that can enable the visualization and understanding of metastasis at each step of the process.