Dendritic cells (DCs) constitute a specialised system of antigen-presenting cells with a high capacity to induce and to modulate the immune response against microbial, tumour and self-antigens. New techniques to generate large amounts of DCs together with the molecular identification of human tumour-associated antigens (TAA) have opened new ways for antigen-specific cancer immunotherapies. DCs loaded either with TAA-derived MHC class I-specific synthetic peptides or with whole tumour cell preparations have been used in numerous clinical trials evaluating the efficacy of DCs in patients with cancer. However, the disadvantages of DCs pulsed with synthetic peptides from TAA include the uncertainty regarding the longevity of antigen presentation, the restriction by the patient's haplotype and the relatively low number of known MHC class I and in particular of MHC class II helper cell-related epitopes. Whole tumour cell preparations are difficult to standardise, and they depend on the availability of tumour cells. Thus the utilisation of viral vectors genetically modified to express TAA for the ex vivo transduction of DCs is an attractive alternative to achieve a MHC I- and MHC II-restricted presentation of tumoural antigens. To induce protective anti-tumoural immune response an increasing number of modified viral vectors have been used to transduce DCs. Although high transduction efficacies were reported for several viruses, analysis of the interaction of viral vectors with DCs has revealed several viral mechanisms that interfere with main functions of DCs, dampening somewhat the initial optimism in the field of DC transduction. However, promising results with different vectors have been achieved. In this review we summarise available data and discuss advantages and drawbacks of currently available vectors.