Effects of lowering the proposed top-concentration limit in an in vitro chromosomal aberration test on assay sensitivity and on the reduction of the number of false positives

For the in vitro chromosomal aberration (CA) test, the proposed top-concentration limit will be reduced to '10mM or 2mg/mL' (whichever is lower) in the draft revised OECD (r-OECD) test guideline (TG) 473, down from '10mM or 5mg/mL' in the current OECD TG, which was adopted in 1997 (1997-OECD). It was previously reduced to 1mM or 0.5mg/mL in the International Conference of Harmonization (ICH) S2 (R1) guideline for pharmaceuticals. Reduction of the top-concentration limit is expected to reduce the number of false or misleading positives. However, this reduction may affect the sensitivity or specificity to predict rodent carcinogenicity. Thus, the effect of a reduction in the top-concentration limit on sensitivity and specificity was investigated by use of a dataset on 435 chemicals obtained from the 'Carcinogenicity and Genotoxicity eXperience' (CGX) database (267 CA-positives and 168 CA-negatives; 317 carcinogens and 118 non-carcinogens) where three TGs (i.e., 1997-OECD, r-OECD and ICH) were applied. The application of the r-OECD TG did not affect the sensitivity (63.1%) or specificity (59.3%) against carcinogenicity, compared with the 1997-OECD TG (sensitivity 63.1%, specificity 59.3%). However, the application of the ICH TG had certain effects, i.e., a decrease in sensitivity (45.4%) and an increase in specificity (72.9%). A change in the number of CA-positives by the application of each TG was also investigated by use of 124 CA-positives from the Japanese Existing Chemical (JEC) database. The application of r-OECD TG showed a small reduction in CA-positives, but the ICH TG reduced this number by approximately half. More than half of the CA-positives had a molecular weight <200. These results suggest that the r-OECD TG will not affect the sensitivity or specificity for the detection of rodent carcinogens, indicating the usefulness of the guideline. However, nearly no improvement with respect to a reduction in the number of false positives should be expected.