Replication-selective oncolytic viruses are being developed as novel, targeted anticancer agents (virotherapy) (1,2). Viruses have evolved to infect cells, replicate, induce cell death, release viral particles, and, finally, to spread in human tissues. Replication in tumor tissue leads to amplification of the input dose at the tumor site, whereas a lack of replication in normal tissues can result in efficient clearance and reduced toxicity. Revolutionary advances in molecular biology and genetics have led to a fundamental understanding of both the replication and pathogenicity of viruses and carcinogenesis. These advances have allowed novel agents to be engineered to enhance their safety and/or their antitumoral potency (2,3). One approach to engineering adenovirus selectivity has been to complement loss-of-function mutations in cancers with loss-of-function mutations within the adenovirus genome. Many of the same critical regulatory proteins that are inactivated by viral gene products during adenovirus replication are also inactivated during carcinogenesis (4–7). Because of this convergence, the deletion of viral genes that inactivate these cellular regulatory proteins can be complemented by genetic inactivation of these proteins within cancer cells (8,9). This concept was proven by the E1B55kD deletion in the dl1520 mutant virus (Onyx-015), which is selective for cells lacking functional p53 (8) enabling viral replication in the majority of human cancers. Clinical trials with dl1520 ultimately proved the selectivity and relative safety of this approach (i.e., refs. 10–12). However, more efficacious adenoviruses are needed.