beta-Carotene and Retinol Efficacy Trial is a nationwide chemoprevention trial that recruited 18,314 high-risk individuals to test the effect of supplemental beta-carotene and retinol on lung cancer incidence. In this report, we conducted a prospective nested case-control study of the association between serum carotenoids, retinoids, and tocopherols on both lung and prostate cancer incidence. Prerandomization serum samples were selected from 278 lung cancer cases and 205 prostate cancer cases, and 483 controls matched by high-risk population, study center location, age, sex (lung cancer only), smoking status, and year of randomization. Carotenoids, retinoids, and tocopherols were analyzed by high-performance liquid chromatography. Endpoints were confirmed by pathology review (lung cancer) or review of the pathology report (prostate cancer). In the control-only population, there was a significant association between tobacco use and serum micronutrient concentration. Current smokers compared with former smokers had lower mean levels of all of the micronutrients tested with zeaxanthin, beta-cryptoxanthin, alpha-carotene, alpha-tocopherol, retinol, and retinyl palmitate reaching statistical significance at P = 0.05. In the overall population, the mean serum concentrations of all of the micronutrients except gamma-tocopherol were lower for lung cancer cases than controls. Statistically significant trends across quartiles were observed in lutein (P = 0.02), zeaxanthin (P = 0.02), and alpha-tocopherol (P = 0.03). The carotenoid findings in the overall population were because of the strong inverse association between serum micronutrients and lung cancer in females. Statistically significant odds ratios (ORs) comparing 4(th) to 1st quartiles in the female population were seen in lutein [OR, 0.31; confidence interval (CI), 0.13-0.75], zeaxanthin (OR, 0.31; CI, 0.12-0.77), and beta-cryptoxanthin (OR, 0.34; CI, 0.14-0.81). For prostate cancer, mean serum concentrations were lower in cases for all of the nutrients except alpha-carotene. Only for alpha-tocopherol (P(trend) = 0.04) were the findings statistically significant. There was no statistically significant association between serum carotenoids and prostate cancer. Our findings provide additional support for the association between physiological levels of dietary micronutrients and cancer incidence.