Osteosarcoma, hemangiosarcoma, and glioblastoma multiforme are three types of incurable cancers that are responsible for reduced quality of life and significant mortality in dogs. The processes that control the clinical behavior of these tumors is not well understood, but recent research suggests that a specialized group of cells called “cancer stem cells” (or CSCs) might play important roles. CSCs are important in the development, growth and spread of various malignant tumors of humans, so it would not be surprising if they also contributed to canine cancers. However, the definitive existence of CSCs in canine cancers remained to be proven. Defining these cells and understanding their basic biology will be necessary to develop cures for cancer, so we developed a unique, widely applicable method to enrich for cells that behave like CSCs from canine tumors. This method allowed us to make direct comparisons among CSCs from osteosarcoma, hemangiosarcoma, and glioblastoma to find unique or shared targets for prevention and therapy. Our objective was to show that the presumed CSCs are indeed uniquely capable of forming tumors using the gold standard assay of “xenotransplantation at limiting dilution”. We confirmed the utility of our method and demonstrated the enrichment of CSC within our culture system. Additionally, we demonstrate that our system will enrich for CSC in tumors where they are infrequent, but provide minimal enrichment for those (aggressive) tumors where a CSC has already become enriched. This provides a new and intriguing possibility to use our culture system to evaluate tumors for the presence of CSC and potentially make therapeutic decisions based on the result. This work also helped provide critical support for further groundbreaking genetic analyses of CSCs that will help us reach the promise of targeted therapies to reduce morbidity and mortality and improve outcomes for dogs with these otherwise deadly cancers.