The feasibility of using adenoviruses for gene therapy has been under close scrutiny recently, as it has become clear that significant toxicity can result from the strong immune response created by intravenous administration of large doses of first generation adenovirus vectors. This suggests that other vectors could be more useful for treatment of metabolic and hereditary disease, where widespread transduction is often necessary for effective gene replacement, and the viability of target cells is important. However, promising recent results in human cancer trials have confirmed that adenoviruses can be very useful in oncology. For cancer treatment, the unparalleled transduction efficacy of adenovirus in dividing and dormant cells is a major benefit. As the goal in cancer gene therapy is to kill infected tumour cells, long-term transgene expression is not necessary. In addition, the immune response generated against infected cells could be useful for eradicating uninfected tumour. Importantly, more than 670 cancer patients have been treated with adenovirus intratumorally, intra-arterially, intraperitoneally and intravenously with very manageable adverse effects and no unexpected severe or lethal toxicity. Currently, the most promising approaches are based on replication-competent agents that allow efficient tumour penetration because of their capacity for tissue-specific replication. In addition to transcriptional control, it is becoming clear that targeting is necessary for efficient tumour transduction and less infection of normal tissues. Exciting results are anticipated when the first selectively replicating targeted adenoviruses go to clinical trials. In conclusion, intense gene therapy and virological research have suggested that while other vectors could be more useful for treatment of hereditary disease, adenoviruses are highly promising and safe agents for oncology, as suggested in a number of early phase clinical trials.