Purpose: Current animal tumor models are inadequate for the evaluation of toxicity and efficacy of conditionally replicative adenoviruses. A novel model system is needed that will provide insight into the anticipated therapeutic index of conditionally replicative adenoviruses preclinically. We endeavored to show a novel model system, which involves ex vivo evaluation of conditionally replicative adenovirus toxicity and therapeutic efficacy in thin, precision-cut slices of human primary tumor and liver. Experimental Design: The Krumdieck thin-slice tissue culture system was used to obtain and culture slices of tumor xenografts of ovarian cancer cell lines, human primary ovarian tumors, and human liver. We determined the viability of slices in culture over a period of 36 to 48 hours by ([3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxphenyl-2-(4- sulfophenyl)-2H-tetrazolium, inner salt)]) (MTS) assay. In vitro Hey cells, slices of Hey xenografts, and human ovarian tumor or human liver slices were infected with 500vp/cell of either replication competent wild-type adenovirus (Ad5/3wt), conditionally replicative adenovirus (Ad5/3cox-2), or the replication deficient adenovirus (Ad5/3luc1). At 12-, 24-, and 36-hour intervals, the replication of adenoviruses in these slices was determined by quantitative reverse transcription-PCR of adenoviral E4 copy number. Results: Primary tumor slices were able to maintain viability for up to 48 hours after infection with nonreplicative virus (Ad5luc1). Infection of Hey xenografts with Ad5/3cox-2 showed replication consistent with that seen in Hey cells infected in an in vitro setting. Primary tumor slices showed replication of both Ad5/3wt and Ad5/3cox over a 36-hour time period. Human liver slices showed replication of Ad5/3wt but a relative reduction in replication of Ad5/3cox-2 indicative of conditional replication "liver off" phenotype, thus predicting lower toxicity. Conclusions: The thin-slice model system represents a stringent method of ex vivo evaluation of novel replicative adenoviral vectors and allows assessment of human liver replication relative to human tumor replication. This is the first study to incorporate this system for evaluation of therapeutic efficacy and replicative specificity of conditionally replicative adenoviruses. Also, the study is the first to provide a valid means for preclinical assay of potential conditionally replicative adenovirus-based hepatotoxicities, thus providing a powerful tool to determine therapeutic index for clinical translation of conditionally replicative adenoviruses.