Comparison of the EU T25 single point estimate method with benchmark dose response modeling for estimating potency of carcinogens

The T25 single-point estimate method of evaluating the carcinogenic potency of a chemical, which is currently used by the European Union (EU) and is d enoted the EU approach, is based on the selection of a single dose in a chr onic bioassay with an incidence rate that is significantly higher than the background rate. The T25 is determined from that single point by a linear e xtrapolation or interpolation to the chronic dose (in mg/kg/day), at which a 25% increase in the incidence of the specified tumor type is expected, co rrected for the background rate. Another method used to obtain a carcinogen ic potency value based on a 25% increase in incidence above the background rate is the estimation of a T25 derived from a benchmark dose (BMD) respons e model fit to the chronic bioassay data for the specified tumor type. A co mparison was made between these two methods using 276 chronic bioassays con ducted by the National Toxicology Program. In each of the 2-year bioassays, a tumor type was selected based on statistical and biological significance , and both EU T25 and BMD T25 estimates were determined for that end point. In addition, simulations were done using underlying cumulative probability distributions to examine the effect of dose spacing, the number of animals per dose group, the possibility of a dose threshold, and variation in the background incidence rates on the EU T25 and BMD estimates. The simulations showed that in the majority of cases the EU T25 method underestimated the true T25 dose and overestimated the carcinogenic potency. The BMD estimate is generally less biased and has less variation about the true T25 value th an the EU estimate.