Cadmium is a metal that is toxic to humans, and the major source of cadmium exposure in the non-smoking general population is diet. To identify major food sources and lower exposure from diet, an accurate estimate of dietary cadmium intake is needed. Hence, the objectives of this study are to develop a method to assess dietary cadmium intake using a biomarker measurement and to improve the estimation of dietary cadmium intake when using a food frequency questionnaire (FFQ). In a random sample of an Italian population, we collected dietary habits by FFQ and measured cadmium in foods and beverages. These data were used to compute the estimated weekly dietary intake (WDI) of cadmium (mu g) by kilogram (kg) of body weight (bw) (WDIFFQ). We also measured fasting serum cadmium levels by inductively-coupled plasma mass spectrometry. We used these data to develop a model for the estimation of the biomarker-derived dietary cadmium intake (WDIbio). In the 51 subjects recruited, the median level of serum cadmium was 0.041 mu g/L (interquartile range (IQR): 0.030-0.054). The median WDIFFQ and WDIbio were 1.34 mu g/kg bw/week (IQR: 0.86-1.70) and 0.72 mu g/kg bw/week (IQR: 0.55-1.11), respectively. The correlation between the two estimates was low-to-moderate (r = 0.291). In exploratory analyses, the correlation was slightly higher in women and participants ages <50 years, and markedly higher in participants with body mass index <25 kg/m(2) and smokers. Our approach allows for the dietary contribution to be isolated from the overall cadmium exposure measured with a biomarker; the estimated dietary cadmium intake was roughly similar to that estimated using the FFQ, especially in select subgroups. Future refinements to the biomarker-derived dietary cadmium intake approach should take into consideration additional sources of cadmium exposure, as well as factors affecting its absorption and metabolism.