We examined the association of breast cancer risk as assessed by mammographic density with dietary patterns in a case-control study nested within the Hawaii component of the multiethnic cohort. This analysis included 3512 prediagnostic mammograms from 1250 premenopausal and postmenopausal women. Mammographic density of the breast was quantified by a computer-assisted assessment method. All study participants completed a validated Food Frequency Questionnaire at entry into the multiethnic cohort. Factor analysis was performed to extract dietary patterns. We computed covariate-adjusted mean percentage densities by quartiles of factor scores using mixed models and weights to account for breast cancer prevalence in the population. Women in this study were primarily of Caucasian, Japanese or native Hawaiian ethnicity. The mean percentage density was 33%. The following three dietary patterns were identified: 'fat and meat', 'vegetables' and 'fruit and milk'. The differences in densities across quartiles of any patterns were relatively small and did not reach statistical significance. Women who had higher scores in the 'fat and meat' pattern had higher densities than those with lower scores (34.8 vs. 32.3%, P for trend=0.21). The 'vegetables' pattern and deep-yellow vegetable intake were weakly inversely associated with densities only among Japanese women (P for trend=0.13 and 0.03, respectively). Our findings indicated that meat and plant-based nutritional patterns are related to breast density. The influence of diet on mammographic densities, however, appears to be fairly small. Analyses of both dietary patterns and single dietary factors seem equally informative in examining dietary associations with percentage density.