To evaluate the relationship between carotenoid concentrations in serum and breast tissue, we measured serum carotenoid concentrations and endogenous carotenoid levels in breast adipose tissue of women with benign breast tumor (n = 46) or breast cancer (n = 44). Before extraction, serum was digested with lipase and cholesterol esterase, and breast adipose tissue was saponified. Serum and tissue carotenoids were extracted with ether/hexane and measured by using HPLC with a C30 column. Serum retinoic acid was extracted with chloroform/methanol and measured using HPLC with a C18 column. There were no significant differences in serum carotenoids [lutein, zeaxanthin, cryptoxanthin (both alpha- and beta-), alpha-carotene, all-trans beta-carotene, 13-cis beta-carotene and lycopene], retinoids (retinol, all-trans and 13-cis retinoic acids), and alpha- and gamma- tocopherol concentrations between benign breast tumor patients and breast cancer patients. A substantial amount of 9-cis beta-carotene was present in adipose tissue and was the only carotenoid that had a significantly lower level in benign breast tumor patients than in breast cancer patients. Correlations between carotenoid concentrations in serum and in breast adipose tissue were determined by combining the data of the two groups. Concentrations of the major serum carotenoids except cryptoxanthin showed significant correlations with breast adipose tissue carotenoid levels. When the concentrations of serum carotenoids were adjusted for serum triglycerides or LDL, correlations between serum carotenoid concentrations and breast adipose tissue carotenoid levels markedly increased, including that of cryptoxanthin (P <0. 001). The strong correlation between serum carotenoid concentrations and endogenous breast adipose tissue carotenoid levels indicate that dietary intake influences adipose tissue carotenoid levels as well as serum concentrations, and that adipose tissue is a dynamic reservoir of fat-soluble nutrients.