Is the extrapolation of carcinogenicity results from animals to humans always feasible?

The extrapolation of carcinogenicity data obtained in animals to humans is not always feasible due to various factors recently evidenced. Among these factors, comparative toxicokinetic, toxicogenomic and genetic susceptibility approaches have more and more importance also in classification of carcinogenic potential of chemicals to humans (either proven, probable, possible carcinogens for humans or no safety concern for humans). This is the base to perform a quantitative cancer risk assessment whose results could even be null when the only tumors observed in small rodents are really sex/strain/species-specific tumors. It would be desirable that companies allow sharing of the main information about the environmental and health effects of their patent molecules. In such a way the confidentiality will be skipped at least for these relevant information on new (and better) molecules.