Problems in Validating the Ames Assay as a Predictor of Human Environmental Carcinogenic Risk

Abstract

<p>The Ames Assay is a battery of tests which measures the mutagenicity of chemical substances. In order for its results to be relevant to human health, Ames Assay mutagenicity must relate to human cancer in a biologically credible and comprehensive way. It must also correlate with other measures and attributes of carcinogenicity. The evidence presented demonstrates its biologic credibility, but illustrates the improbability of its being a comprehensive measure of all attributes of carcinogenesis. Agreement between the Ames Assay and animal cancer studies, other short-term bioassays, human carcinogens, and chemical structures predictive of carcinogens is presented. The range of possible predictive accuracy for the Ames Assay is calculated from these agreement evaluations. A hypothetical model is developed to test whether the Ames Assay is a sufficiently valid predictor of human carcinogenesis regardless of the extremes of this range. It is not sufficiently valid in the lower half of the range. So more work must be done to define precisely its accuracy.</p> <p>The problems of applying the range of accuracy of a laboratory test of pure substances to biologic and environmental samples are explored. Unsolved problems in collection and analysis of airborne samples are identified. A study design is presented which minimizes the chance of bias in collecting airborne samples. The limitations of the laboratory procedure and the problems of sample collection are brought together in two continuous flow schemes. Together these demonstrate the unsolved biologic and biostatistical problems in validating the Ames Assay.</p>