Design, analysis and reporting of tumor models

An overview of preclinical in vivo tumor growth studies The analysis of in vivo tumor growth is a powerful tool for studying the effects of potential cancer therapies and understanding inherent variability in a given animal subject or tumor line, which can be used to guide effective study design. In a typical tumor growth study, human tumor cells or tumor fragments are inoculated into immune compromised rodents, typically mice. The tumor bearing animals are then blocked at a given time point based on tumor volume and randomized into different groups that will receive a potential therapeutic or a matched vehicle control. Tumor growth is then assessed periodically throughout the course of the experiment, normally by use of caliper measurement or by more complex imaging techniques depending on the site and nature of the tumor under study. The resulting data set consists of a series of tumor volumes for each individual animal over time, which can be analyzed to determine whether the potential cancer therapy affects tumor growth compared to control. A standard practice to determine the therapeutic effect of a test agent is to determine if there are meaningful differences between tumor growth in control vs. treatment groups and thus significance tests are often performed. There are, however, many ways to test statistical hypotheses and not all are equally well suited for a given study and depend on research goals and tolerance for false positives (Type I errors) or false negatives (Type II errors). Likewise, most statistical tests require certain assumptions about the data to be maintained; if they are contravened, the analysis and conclusions of the study would be invalid.