Abstract:
|
The single induction test procedure for modeling multiple tumor data by Kodell and Chen (2001) is extended to a test for tumor multiplicity in multiple induction experiments. The new test can detect overall differences in the distribution of observed tumors between two groups, as well as isolate whether the detected overall difference is due to the difference in the distribution of the number of induced tumors and/or the difference in the distribution of the time to tumor observation. The behavior of the proposed test was studied via Monte Carlo simulations. To simulate realistic data, we generated data from parameters that were similar to some of the parameter estimates from previous experiments. In a typical photococarcinogenicity experiment, a group of male and female mice are exposed to no light or two doses of light following either no treatment or treatment with a test agent for forty weeks, and the occurrence, size, and multiplicity are recorded weekly for each mouse. Results of the proposed test were compared to those of the logrank test, the negative binomial test, and Dunson's test (2000). Monte Carlo simulation studies showed that parameters were reasonably estimated.
|