Purpose: Despite the success of conditionally replicating adenoviruses in tumor models, clinical success has been limited when they are used as a single modality agent. Overcoming the disparity in efficacy between in vivo animal models and human use is a key hurdle for better conditionally replicating adenovirus therapy in humans. We endeavored to identify biological blocks to adenoviral infection and replication in tumor cells. Experimental Design: We hypothesized that the differences in adenoviral replication between ovarian cancer cell lines and patient tumor samples are the result of a block in viral RNA transcription. To test this hypothesis, established ovarian cancer cell lines and purified patient ovarian cancer cells were infected with wild-type adenovirus. RNA for early adenoviral genes E1A and E1B as well as the late transcripts for fiber and hexon were measured using real-time PCR. Results: Established ovarian cancer cell lines treated with wild-type virus had a lower E1A:E1B ratio than the patient samples. Additionally, the levels of fiber and hexon relative to E1A were also decreased in the patient samples compared with the established cell lines. These findings were consistent with an early- to late-phase block in the adenovirus replication cycle. Conclusions: These data suggest that the biology of abortive infection in the patient samples may be linked to a defect in the production of early and late viral transcripts. Identification of factors leading to abortive infection will be crucial to understanding the low viral replication in patient samples.