Epidemiological studies suggest that a high vegetable diet may reduce risk for breast cancer and may also improve prognosis after the diagnosis of breast cancer. Circulating carotenoids may serve as a biomarker of vegetable and fruit intake, although several factors affect their bioavailability from food sources and may influence concentrations. One purpose of this study was to identify factors predictive of serum carotenoid, retinol, and alpha-tocopherol concentrations in 79 postsurgically resected breast cancer patients at enrollment and at 12 months in a feasibility study of a high vegetable, low fat diet intervention to reduce risk for cancer recurrence. Another purpose was to identify variables associated with change in these serum concentrations 12 months after randomization into control and intervention groups. The diet intervention (versus control) group had significantly greater increases in carotenoid intakes (P < 0.03) and significantly greater increases in serum concentrations of lutein, alpha- and beta-carotene, lycopene, and retinol (P < 0.04). Stepwise multiple regression revealed the level of dietary intake to be predictive of most serum carotenoid concentrations at baseline and 12 months, with additional associations between selected micronutrient concentrations and serum cholesterol, body mass index, age, percentage of energy intake from fat, and alcohol intake also observed at these time points. Intervention group change in serum carotenoid concentrations was inversely associated with baseline level, age, and change in serum cholesterol concentration and positively associated with change in carotenoid and alcohol intake. Circulating carotenoid concentrations are responsive to a high vegetable diet intervention, which also included reduced dietary fat and increased fiber intakes, to reduce risk for breast cancer recurrence.