Rater agreement and utility of the mutagen-induced chromosome damage assay

Chromosomal damage in peripheral blood lymphocytes induced by short-term in vitro exposure to the cytotoxic antibiotic bleomycin was first described in 1983 and proposed as a phenotypic assay for chromosome instability. This assay was subsequently described as potentially useful in assessing an individual's risk to environmental carcinogens in 1989. Since 1995 numerous published studies have used this assay to assess risk for cancer in the aerodigestive tract, particularly lung cancer, in various ethnic populations. Odds ratios up to 8.5 have been reported for individuals deemed "mutagen sensitive" (defined as > or = 1 chromatid break per metaphase averaged in 50 metaphases analyzed). While this phenotypic assay is appealing for lung cancer risk assessment it has not been reproduced by other investigators. Because of our interest in lung cancer biology, epidemiology, and genetics, we sought to independently assess the rater agreement of this assay. We found that 1) the assay is laborious to conduct (8 hours of labor) and relatively expensive (> $100), yet reducing the number of metaphases from 50 to 20 produced a reliable, less expensive, and less laborious test; and 2) the rater agreement of individual metaphase readings is poor, but agreement for a summary measure is high.