Historical perspective of the cancer bioassay.

For more than half a century the rodent bioassay has been used in an attempt to predict the potential of chemicals to cause cancer in humans. High mortality and animal husbandry problems led to a more statistically correct study that utilized 50 or more animals per group, cost 1 to 3 million US dollars, and required 3 to 6 years to complete. With an increased understanding of the carcinogenic process came the realization that chemically induced cancer in rodents may not be predictive of carcinogenic potential in humans. This paper describes the odyssey of the bioassay and recommends design improvements that include the incorporation of new molecular methods, the development of data to describe a plausible mode of action, pharmacokinetic and dynamic data for cross-species extrapolation, relevant dose levels consistent with human exposure scenarios, reduction in the number of animals used, and a harmonized international testing strategy for the acquisition of data useful in risk assessments.