Thyroid follicular cell carcinogenesis: results from 343 2-year carcinogenicity studies conducted by the NCI/NTP.

The National Toxicology Program data base on 343 mouse and rat carcinogenesis studies was reviewed to determine the frequency of and relationship between hyperplastic and neoplastic follicular lesions of the thyroid gland. The frequency of chemically related lesions in the thyroid was also compared to neoplastic lesions in the liver to investigate a possible correlation. The percentage of studies observed to have positive or equivocal chemically related thyroid proliferative lesions was rats: male, 14%, female, 11%; mice; male, 8%; female, 9%. When positive in one sex for a given chemical, there was a 60-80% chance of it being positive in the other sex of the same species, although interspecies correlation was not as strong. Thyroid follicular cell neoplasia without hyperplasia was uncommon in mice but was common in rats. Chemicals that caused thyroid proliferative changes were more likely (P less than 0.05) to produce liver neoplasms (both within and between species) than were chemicals causing no thyroid changes. However, this correlation was far from perfect, with many chemicals producing thyroid proliferative lesions, but not liver neoplasms and vice versa. This suggests that universal correlations are not supportable by the data and that individual chemicals should be evaluated on a case-by-case basis.