Can carcinogenic potency be predicted from in vivo genotoxicity data?: a meta-analysis of historical data.

Genotoxicity is generally a parameter used for hazard identification, however, the applicability of using in vivo genotoxicity tests for hazard characterization has never been thoroughly investigated in a quantitative manner. Genotoxicity assays could be useful for the determination of cancer potency parameters given that genotoxicty tests measure mutations and/or chromosomal aberrations which are strongly associated with carcinogenesis. A detailed literature survey was performed in search for dose-response data in various in vivo genotoxicity and carcinogenicity studies. The benchmark dose (BMD) approach was applied using the dose-response modeling program PROAST. Dose-response data were available from 18 compounds in the micronucleus assay (MN), the in vivo transgenic rodent mutation assay (TG) and the comet assay, and their BMD(10) values were compared to the BMD(10) from carcinogenicity studies in mice. Of the 18 compounds, 15 had acceptable dose-response data from the MN and the TG, but only 4 from the comet assay. A major limitation in our analysis was the lack of proper dose-response studies using the recommended protocols. Nevertheless, our findings are promising because even with these suboptimal studies, a positive correlation was observed when the lowest BMD(10) from the genotoxicity tests (MN and TG) was compared to the tissue-matched carcinogenicity BMD(10) . It is evident that more compounds need to be analyzed with proper dose-response schemes to further validate our initial findings. Experimental designs of genotoxicity assays need to shift from focusing only on hazard identification where positive and negative results are reported, to a more quantitative, dose-response assessment.