|dc.description.abstract||Cancer is the second leading cause of death in humans worldwide and the most common cause of death for humans in developed countries. Cancer is also the leading cause of mortality in dogs. Animal models, primarily genetically engineered mice, have been used historically in research aimed at discovering causes and treatments for human cancer types, but the canine model has been underutilized in this research area. Though the field of canine cancer epidemiology is relatively new, it has great potential to produce answers to research questions pertaining to cancer prevention, development, and treatment relevant to both dogs and humans. In fact, the canine spontaneous tumor model is actually a better model for use in human cancer epidemiology research than other animal models or even human populations. This is due to the fact that dogs spontaneously produce many different types of tumors that are molecularly indistinguishable from human tumors. Canine DNA shares a large amount of ancestral sequence with human DNA, but dogs have greater genetic homogeneity - even across breeds - than do humans, which simplifies disease mapping at the genomic level. Dogs live in the same environments as humans, too, so they share many similar exposures to environmental factors that may contribute to the development of cancer. Tumors in dogs progress at a rapid rate as compared to humans and many tumor types that are rare in humans occur frequently in dogs. These facts, when considered along with the existence of an accelerated aging process in dog, support how use of the canine spontaneous tumor model will allow us to gain a greater understanding of genetic and environmental contributions to human disease and do so at a rapid pace. The primary aims of my dissertation were to examine the current body of evidence produced through canine cancer epidemiology research, produce new research using study designs similar to those used in human cancer epidemiology research, and show how we can advance knowledge of cancer risk and pathogenesis in both fields using the canine spontaneous tumor model. • Through utilization of systematic review methodology I was also able to identify a lack of consistency in study design and statistical methods used in veterinary cancer research even when exploring the same research question. • Then, with a study population of nearly 68,000 dogs admitted to hospitals across the United States over a 20-year time frame, I was able to show there are several breed-specific and hormone-dependent risks associated with development of lymphoma, many of which had not been previously reported, likely due to the use of small samples sizes in veterinary cancer research. • I also discovered differences in the geographic distribution of dogs diagnosed with two different subtypes of lymphoma in the US. This implies molecular characterization of some cancers, which is commonly done in human epidemiologic research, may be a necessary component of future veterinary epidemiologic research in order for us to truly identify risks for disease occurrence. • Lastly, I utilized a study population composed only of Bernese mountain dogs, a breed known to be predisposed to developing a cancer type that is rare in humans, to investigate exposure variables associated with disease outcome. The results of this project suggested a mechanism of disease pathogenesis not previously reported in either the veterinary or the human literature. As a body of work, these individual studies contribute to advancing the concept of using canine spontaneous tumor models in lieu of using other animal models for comparative research. In addition, the canine model can be superior to use of human subjects in many instances, including when determining the etiology of rare cancer types or when determining the pathogenetic basis of disease. In order to continue positively contributing to the field of cancer epidemiology, veterinary epidemiologists must increase the rigor with which we are conducting studies, report research in a transparent manner by conforming to accepted reporting guidelines, and ensure we are investigating appropriate research questions.