The formation of carcinogen-DNA adducts within the respiratory epithelium is thought to be a critical factor in the induction of lung cancer from tobacco smoke. A reliable surrogate measure of carcinogen damage to the lung would be of great value in molecular epidemiological studies of cancer risk. The validity of measurements of DNA adducts formed from hydrophobic aromatic hydrocarbons in peripheral blood mononuclear cells (MNCs) was investigated by comparing the levels of aromatic DNA adducts detected in lung tissue from 31 lung cancer patients with those detected in MNCs from the same individuals using the 32P-postlabeling assay. The associations of smoking history and intake of dietary antioxidants with adduct levels also were assessed. Tissue-specific, as well as common DNA adducts were detected in lung and blood; total MNC adduct levels were highly correlated with total lung adducts. After smoking cessation, adduct levels appeared to decay in both tissues at similar rates. Multivariate analyses (Poisson regression modeling) indicated that dietary antioxidant intake (carotenoids, vitamin A, and retinol) modified the levels of aromatic DNA adducts in both the lungs and blood. Of all models tested, the optimal one for predicting lung adduct levels included the measure of blood MNC adduct levels only. Therefore, blood MNCs are a valid surrogate tissue for estimating the burden of DNA adducts in respiratory tissue in molecular epidemiological studies.