However, these models tend to have powerful driving oncogene mutations and other very specific genetic changes that may limit the scope of their applicability to clinical oncology. For the PDX model, each patient’s cancer is unique and represents countless genetic changes that accrued over the time the cancer developed. These cells are not intentionally transformed or forced to overexpress specific proteins like cell lines and transgenic mouse models, respectively. Ideally this means PDXs should more closely resemble the primary tumor. As described in the introduction, many groups have validated the utility of this system, including the capacity for PDXs to recapitulate the metastatic potential of the primary tumor. Importantly, Tentler et al. recently reviewed the translational track record for PDXs in LY2835219 CDK inhibitor oncology drug development for a wide range of cancers, and the future for PDXs in the development of predictive biomarkers for clinical trials appears promising. PDXs can be injected into mice in either an orthotopic or heterotopic manner. The PDXs that we have generated, utilized and described throughout this paper were implanted in a heterotopic fashion since HNCs were grown subcutaneously in the mice. Other groups studying pancreatic, lung, and renal cancers have also utilized subcutaneous tumor implantation for their work. Subcutaneous tumors allow for easier access for size measurements by calipers during therapeutic studies. Moreover, tumors can be grown to a much larger volume subcutaneously as compared to orthotopic sites before causing harm to the mice. In this manner, subcutaneous tumors allow for greater amplification of this valuable tissue, which can then be harvested for histological as well as molecular analyses and for passaging to new mice. Orthotopic implantation, whereby tumor cells are injected at the site of origin, represent another viable manner to propagate and study PDXs. This technique has been described for breast, brain, and HNCs. It is thought that this model allows tumors to develop in a more representative microenvironment. Indeed this system is especially useful for studying metastasis. However, in particular for the head and neck region, only small tumor volumes can be generated before causing harm to the mice. Qiu et al. documented that after two weeks mice with oral PDXs required modified diets due to difficulty eating. Owing to these small tumor sizes, it is difficult to propagate HNC PDXs grown orthotopically. It is clear that heterotopic and orthotopic PDXs both have unique advantages and disadvantages for modeling human cancer. We believe that each of the oncologic models discussed as well as others that are available can and should play a role in future research. Indeed it is important to define the spectrum of utility for every system and determine how each can be maximized to answer specific questions. For example, novel, targeted therapeutics could be initially employed in transgenic mice that are made to overexpress the targeted protein. This information could then translate into screening PDXs for the protein of interest, assessing their relative response to the novel drug, and identifying predictive biomarkers. Even within a specific model system, such as the PDXs, it is important to employ the spectrum of available methods including both heterotopic and orthotopic tumors. Orthotopic tumors appear superior for studying metastasis, especially in the head and neck region, while subcutaneous tumors allow for easier evaluation of therapeutic response with less harm to the mice. In this manner, we feel the strengths of each model can be exploited to advance the treatment of human cancers. Although our methods were carefully planned and executed, there are, of course, limitations to this work. First, we did not carry out the full experimental protocol on fresh patient tissue owing to the small quantity of tumor we receive from the OR. Therefore, we would not have been able to make the number of injections necessary to carry out this work. Instead we performed a smaller experiment with three new PDXs to determine if they retained viability up to 24 hours after initial excision from the patient. Next, we did not perform this experiment on each of our PDXs, but we did select two that represent the broad spectrum of growth that we have seen among our cohort of PDXs. While the histological characterization we performed demonstrated no differences based on time or storage medium, molecular changes such as gene methylation or the development of mutations could occur as the time increases between tumor excision and implantation. However, we did not carry out specific analyses to address this potential issue. Finally, we cannot say how much time beyond 48 hours the tumors would remain viable for passaging. Despite the limitations of this experiment, our results are surprising as most researchers in the PDX community, including us, believed strongly in the importance of the time interval to implantation. Therefore, we did not expect the fresh patient tissue to remain viable 24 hours post-resection nor did we anticipate the passaged PDXs would retain equivalent viability across the seven different time points, especially up to 48 hours. These experiments represent the first non-anecdotal evidence in the literature about the relationship between time, storage medium and ultimate PDX development. We hope other investigators can apply our findings to their PDX work in order to maximize the potential of this valuable model system. For example, it is difficult to perfectly monitor the health of immunodeficient mice since they are very fragile and within a short window they can change from clinically healthy to on the verge of death.