To the editor:

In a recent Perspective, Rosenberg and colleagues commented on the need to revise current strategies for therapeutic vaccination against cancer, given disappointing clinical response rates in recent trials1. Among 765 patients treated in 35 trials selected to be representative and encompassing of multiple cancer types (see Table 5 in ref. 1), the cumulative objective response rate was only 3.8%. In trials using poxviruses, peptides and modified tumor cells, clinical responses were seen in 0%, 4.0% and 4.2% of subjects, respectively, and tumor regressions were seen in 7.1% of patients receiving dendritic cell vaccines.

This analysis does not include a study of dendritic cell–based vaccination for non-Hodgkin B cell lymphoma targeting tumor-specific immunoglobulin (idiotype)2,3. In this study, the objective tumor regression rate was 31.6% among 19 patients with disease evaluable for response by standard radiographic criteria (50% decrease in the sum of the products of perpendicular diameters of all lesions; no increase in any lesion), including four durable complete tumor regressions. Although this study had a small sample size like others included in the analysis, the 31.6% objective response rate exceeds all others cited. Given these results, further studies of dendritic cell–based vaccination for lymphoma are now underway, along with several phase 3 randomized studies of idiotype protein vaccines4,5.

Non-Hodgkin lymphomas are largely incurable, vascularized cancers that account for over 23,000 deaths annually in the US6. Rosenberg et al. suggest that easier access to the circulation and the expression of costimulatory molecules by tumor cells may be a key to the clinical responsiveness of lymphoid tumors. However, other factors also deserve consideration, including the nature of the target antigen. One reason the target is important relates to the mechanism of the antitumor effect. Antibodies are believed to be an important component of the immune response to cell-surface targets such as lymphoma idiotype7. The antibody response is prompted through the use of full-length tumor antigen, with Fc receptor-bearing cells possibly serving as effectors7. But durable tumor regressions have also been observed in the absence of measurable anti-idiotype antibodies2,3,4, thereby implicating T cells as critical effectors in some cases. Thus, although the early clinical experience with adoptive T cell therapy for melanoma seems promising in contrast to therapeutic vaccine results reported in melanoma to date8, this approach may not be required to achieve success against all tumor types.

The diverse biology of human cancers is likely to preclude any single tumor type or approach from serving as a dominant paradigm for the development of immunotherapies successful across a broad spectrum of cancers. Rather, the careful tailoring of particular vaccine, adoptive transfer and antibody strategies to particular antigens (and to particular cancers) may translate into the most fruitful outcomes for cancer immunotherapy.