All of these studies taken together highlight key areas that must be addressed in the future in order for the field to continue to move forward. These issues are many, including but not limited to method of delivery of dendritic cells to patients, maturation status of the dendritic cells, and methods of monitoring responses to these vaccines. Each of these requires some comment. Different strategies of immunization were used in these studies. DCs were injected at various times and in various locations, including intradermally/subcutaneously, intranodally, and intravenously. Investigation of the pattern of spread of subcutaneously injected fluorescently labeled DCs in the chimpanzee was studied at the University of Pittsburgh. Although rodent DCs had previously been shown to remain at the site of injection, these immature primate DCs migrated to draining lymph nodes and interact appropriately with T cells for as long as 5 days after administration. Data not shown in the same study reveal that intravenously administered DCs were undetectable in draining lymph nodes. Two groups have undertaken evaluation of route of administration of DCs in humans. The first of these examined migration of immature, indium-111-labeled dendritic cells after RNA-loading in metastatic cancer patients [44]. The DCs were injected either intravenously, subcutaneously, and intradermally. Only DCs injected intradermally were cleared from the injection site with migration to regional lymph nodes. The immunologic significance of these findings is unclear, however. Another study examined this issue by studying prostatic acid phosphatase (PAP) protein-loaded mature DCs injected intravenously, intradermally, and intralymphatically in prostate cancer patients [45]. Regardless of route of administration, T cell responses were induced as measured by proliferation when PBMCs in vitro were stimulated with the PAP protein. Cytokine analysis of the patients revealed that the majority of patients undergoing either intralymphatic or intradermal injection had increases in measurable interferon-gamma but that none of the intravenously-injected patients did. The intralymphatic and intradermal routes thus seem to lead to stronger Th1 responses. But no data was presented regarding the numbers of PAP precursors induced by vaccination nor their specificity/cytotoxicity. Another issue in DC administration that should also affect route of administration is maturation status of the dendritic cells. Many of the studies used immature dendritic cells to immunize patients whereas others used mature cells. A number of studies have demonstrated that maturation signals such as inflammatory cytokines and CD40 ligation lead to down-regulation of antigen processing and up-regulation of the chemokine receptor CCR7, which leads to homing to lymph nodes [46] as well as the MHC molecules, costimulatory molecules, and maturation markers [8,47,48]. In addition, different maturation agents and sequences of addition of these maturation agents may lead to differences in functions of dendritic cells [48-51]. Others have found that injection of immature DCs pulsed with influenza matrix peptide and KLH, and lead to greater numbers of influenza-specific T cells, but these cells had reduced interferon-gamma production and lacked killer activity [52]. Perhaps the most impressive results in a clinical trial, however, were gained by injecting similar cells loaded with melanoma peptides [21]. In addition, sequence of loading and maturation may be important in creating vaccines. One study using CEA peptides and CEA RNA found that optimal T cell presentation occurs when peptides are loaded after maturation with CD40 ligand and when RNA is transfected before maturation with CD40 ligand [53]. As all of these studies reveal, more investigation into the role of DC maturation as well as its timing and sequence is needed. Finally, a multitude of methods to detect response to vaccination have been attempted in all of the above studies. Many use DTH responses, but these may measure CD4 T cells instead of CD8 T cells. The availability of tetramers allows easier quantification of CTL precursors, but they provide no assessment of the function of these T cells. Enzyme-linked immunospot assays allow identification and quantification of numbers of cells producing cytokines such as interferon-gamma when encountering target antigens, but cytokine production may not correlate with tumor cell killing. Chromium release assays or flow cytometric assays for molecules such as perforin may be used to validate killing, but inability of many tumors to grow in vitro precludes direct assessment of tumor cell killing via this method. Other responses in human subjects may also be measured. Some of the cited studies yielded clinical responses that could be measured via physical examination or radiologic study. This is the exception rather than the rule, however. Others have monitored the decrease in serum tumor markers such as PSA or CEA. But these may not correlate directly with tumor burden. Indirect calculation of tumor burden by using quantitative PCR to estimate the number of circulating tumor cells in peripheral blood may be promising in this regard. Despite the lack of consensus as to what constitutes an effective response, most would agree that monitoring of these patients should include measures of both immunologic response and clinical tumor effect. All of this leads to the conclusion that DC-based cancer vaccines have progressed a great deal but that much work still needs to be done. Only rigorous bench top experimentation followed by careful patient selection and vaccine administration, and then by meticulous patient monitoring, will lead to advances in the field.