Effect of discounting certain tumor types/sites on evaluations of carcinogenicity in laboratory animals

It has been suggested that, for mechanistic reasons, certain tumors found in experimental animals should be discounted when evaluating carcinogenic effects. The questioned tumors are: mouse liver, rat thyroid follicular cell, bladder and kidney (male rat), forestomach (mouse and rat, gavage route), and lung (mouse and rat, inhalation of particles). We sought to determine the effects of discounting those tumors on classification of chemicals as carcinogens in animals. We looked at carcinogenicity data for chemicals studied in NCI-NTP bioassays and/or reviewed in IARC monographs and we found that deleting the questioned tumors would have significant impact on evaluations of carcinogenicity in animals. Fifty-six of 234 (24%) chemicals determined to be carcinogenic in the NCI-NTP bioassay program would no longer be considered carcinogenic: 102 (44%) would have weaker evidence of carcinogenic effects. Thirty-three of 361 (9%) chemicals determined by IARC to have "limited" or "sufficient" evidence of carcinogenicity would no longer be considered carcinogenic; 119 (33%) would have weaker evidence of carcinogenic effects. Because such a large number of chemicals currently considered carcinogenic would be affected by categorical deletion of tumors and because we are not aware of data that would justify such categorical deletions, it would be preferable to consider mechanistic justifications for discounting tumors on a case-by-case basis for each individual chemical. Deletion of tumors on a categorical basis has serious implications for regulation of toxic chemicals and for public health.