BACKGROUND & AIMS: We reported an association between cytologic atypia, a reversible biomarker of breast cancer risk, and lower omega-3/omega-6 fatty acid ratio in blood and breast tissue. Our goal was to develop and validate a dietary pattern index in this high-risk sample of U.S. women, and test its capacity to predict incidence in a nested case-control cohort of Canadian women from a randomized trial of a low-fat dietary intervention for primary prevention of breast cancer. METHODS: Food intake was measured by food frequency questionnaire in the U.S. sample (n = 65) and multiple dietary recalls in the Canadian sample (n = 220 cases; 440 controls). Principal component analysis identified a dietary pattern associated with atypia. We measured differences among dietary pattern tertiles in (a) fatty acid composition in blood lipids and breast tissue in the U.S. sample, and (b) risk of breast cancer subtypes in the Canadian cohort. Registered under ClinicalTrials.gov Identifier: NCT00148057. RESULTS: A Modern diet was characterized as consuming more grains, dairy, and sugar and less vegetables, fish and poultry; these women had lower tissue omega-3 fatty acids and higher omega-6 and trans fatty acids. The low-fat intervention increased the likelihood of a Modern diet after randomization. A Modern diet at baseline and post-randomization was associated with estrogen-receptor negative (ER-) breast cancer risk among those at least 160 cm tall. A Traditional diet (the reciprocal of Modern) at baseline was associated with lower ER-positive (ER+) risk in the comparison group, but not the low-fat intervention group. CONCLUSIONS: A Modern diet (high in grains, dairy, and sugar and low in vegetables, fish, and poultry) is associated with ER- breast cancer risk among taller women. Recommending dietary fat reduction may have untoward effects on breast cancer risk.