Lung cancer is a devastating disease that is associated with significant morbidity and mortality. Most patients die of progressive metastatic disease despite aggressive local and systemic therapies [1]. The pathogenesis of lung cancer remains highly elusive due to its aggressive biologic nature and considerable heterogeneity as compared to other cancers. These circumstances substantially impede study of the disease in humans and necessitate the use of experimental models that can be used under more uniform, controlled conditions than those achievable in clinical settings. The development of animal models of lung cancer may aid in our understanding of lung tumor biology and facilitate the development and testing of novel therapeutic approaches and methods for early diagnosis. To this end, animal models should mimic both the genetic alterations found in human lung tumors and their histological characteristics.Humans are one of only a few species that are susceptible to the spontaneous development of lung cancer. Lung tumors in domestic animals were periodically observed by veterinarians, but Livingood's histologic description 100 years ago of a papillary tumor in a mouse [2] initiated the idea of using animals as experimental tools. There are currently several types of animal models that are widely used for experimental lung cancer research. These include chemically induced lung tumors, transgenic mouse models, and human tumor xenografts.A single model system that faithfully reflects the whole process of lung cancer carcinogenesis and progression is unlikely to be developed. Lung cancer models that accurately reflect the different aspects of the disease are necessary to properly investigate its myriad of complexities. Tumorigenesis, proliferation, invasion, angiogenesis, metastasis, prevention, and therapy are all areas where specific models are required to ensure proper experimental design. To reflect the anticipated biological process being studied, model systems may require certain deviations from the human disease. Thus, we should interpret results of studies utilizing model systems with caution and with an appropriate understanding of their limitations. The purpose of this review is to summarize the various lung cancer model systems in use today and to define both their utility and limitations. As Siemann stated, it is best to ª. . . choose the model to address the question rather than force the question on the tumor modelº [3].
General PrinciplesTumor±host interactions, including immunologic effects, vascular and stromal effects, and host-related pharmacologic and pharmacokinetic effects, are poorly modeled in vitro. Animal models to study these areas can be broadly divided into spontaneous or induced tumors, and transplanted tumors. The former group consists of those induced by some extrinsic chemical or carcinogen and animals genetically modified to express genes that lead to lung tumor development. The latter group includes the widely used heterograft and xenograft models. We will briefly describe all of t...