Targeting of oncolytic adenoviruses to tumors can potentially increase their efficacy and safety profile after systemic application. We have developed recently a capsid-modified vector containing the adenovirus serotype 35 fiber shaft and knob inserted into an Ad5 capsid. This Ad5/35 vector infects cells via a coxsackievirus adenovirus receptor-independent pathway. Here we attempted to exploit this new tropism of Ad5/35 vectors for tumor-specific infection. In vitro, the Ad5/35 vector efficiently transduced human breast cancer cells that were refractory to infection with conventional Ad5-based vectors. Additionally, primary mouse hepatocytes were relatively refractory to Ad5/35 infection in vitro or after systemic vector application to mice. In an animal model of breast cancer metastasis, intraportal infusion of MDA-MB435 cells produced multiple hepatic metastases that were surrounded by extracellular matrix and developed blood vessels confined to the tumor stroma. Tail vein injection of a standard Ad5-based vector into tumor-bearing animals resulted in transduction of mouse hepatocytes but not metastases. However, the capsid-modified Ad5/35 vector transduced only approximately 8% of metastases. The metastases that were susceptible to Ad5/35 infection demonstrated blood vessels in close proximity to tumor nests without extracellular matrix separating endothelial and tumor cells. These findings indicate that transduction of liver metastases not only requires tumor-specific tropism but also new strategies to increase accessibility of tumor cells to systemically applied oncolytic adenoviruses.