Concordance of Noncarcinogenic Endpoints in Rodent Chemical Bioassays

Prediction of noncancer toxicologic outcomes in rodent bioassays of 37 chemicals from the National Toxicology Program was evaluated. Using the nonneoplastic lesions noted by NTP pathologists, we evaluate both agreement in toxic lesions across experiments and the predictive value of the presence (or absence) of a lesion in one group for other groups. We compare lesions between mice and rats, male mice and male rats, and female mice and female rats in both short‐term and long‐term bioassays. We also examine whether lesions found in a specific organ in a short‐term test are also found in the long‐term test of the same chemical. We find agreement (concordance) across species for specific lesions, as evaluated by the Kappa statistic, ranging from 0.58 (for concordance of nasal lesions between female mice and rats in long‐term studies) to −0.14 (lung lesions between mice and rats in long‐term studies). Predictive values are limited by the relatively small numbers of observations of each type of lesion. Positive predictive values range from 100% to 0%. Comparing the lesions found in short‐term tests to those found in long‐term tests resulted in Kappa statistic values from 0.76 (spleen lesions in male rats) to −0.61 (lung lesions in female mice). Positive predictive values of short‐term tests for long‐term tests range from 70% to 0%. Overall, there is considerable uncertainty in predicting the site of toxic lesions in different species exposed to the same chemical and from short‐term to long‐term tests of the same chemical.