This project is concerned with statistical methodology issues in the design, analysis, and interpretation of laboratory animal experiments. One investigation demonstrated that while there is a significant correlation between chemical toxicity (as measured by the maximum tolerated dose) and rodent carcinogenicity, this correlation seems limited primarily to gavage studies. Further, the toxicity of a chemical does not seem to be a reliable predictor of whether or not a chemical will be a rodent carcinogen. Another investigation considered the correlation between site-specific carcinogenic effects for chemicals tested in NTP long term studies in rats and mice. For certain sites (e.g., forestomach) there were striking correlations between sexes and species, i.e., chemicals producing these particular tumors tended to increase these neoplasms in both males and females and in both rats and mice. Overall, a chemical producing a site-specific carcinogenic effect in one sex had approximately a 65% chance of producing that same effect in the other sex; similarly, there was approximately a 36% change that a chemical producing a carcinogenic effect in one species would produce that same site-specific effect in the other species. Future research related to this project includes an estimation of the frequency of false positive and false negative outcomes in the interpretation of long term rodent carcinogenicity studies.
