One of the goals of molecular epidemiology is to identify individuals at high cancer risk. Molecular dosimetry of carcinogen exposure and inherited functional polymorphisms of cancer susceptibility genes are the major determinants of an individual's risk. The mutational spectra of the p53 tumor suppressor gene differs in lung cancers from males and females indicating possible gender differences in lung cancer susceptibility that warrant further investigation. The p53 mutational load in nonmalignant tissues may be a molecular dosimetry of carcinogen damage. For example, an increase in 248 mutant p53 cells is found in the colonic epithelial of donors with ulcerative colitis when compared to normal controls. Serum anti-p53 antibodies have been detected in cancer patients and, in some cases, are detected prior to clinical cancer. Case-control studies of human lung cancer are being conducted in the U.S. and the Peoples Republic of China. In the NCI-University of Maryland study we have now found that the combination of genotyping for CYP2D6 A, B and T mutations is predictive of lung cancer risk. The glutathione-S- transferase M1 (GSTM1) and CYP1A1 genetic polymorphisms were studied, but no association of either polymorphism with lung cancer were found. However, racial differences in allelic and genotypic frequencies were found. Urinary mutagenesis was tested in this same population. Recent smoking was associated with mutagenicity and African-Americans were more likely to have mutagenic urine, although when adjusted for recent smoking, the association was not statistically significant. There was no association of the CYP1A1 or GSTM1 genetic polymorphisms urinary mutagenesis. We also have studied lung cancer risk among nonsmoking Chinese women, who have similar cancer rates as U.S. women. In order to explore the molecular epidemiology of lung cancer, we have initiated a study of nonsmoking Chinese women. A total of 168 cases and controls have been collected. Evidence from Japan indicates that the effect of genetic polymorphisms are greatest in nonsmokers. We have hypothesized that this cohort, where there are high cancer rates among nonsmokers, will elucidate genetic predispositions for cancer risk. Thus far, analyses have been completed for CYP1A1, CYP2E1, GSTM1, CYP2D6, APOE, and GST-T have been completed and the data is being analyzed.