Multinomial logistic regression model of the association of breast density and tumor phenotypes with categories of interval cancer, adjusted for age at screening, screening program, and tumor size. The association of breast density with tumor phenotype, adjusted by screening program (categorical), age (continuous), and tumor size (categorical, <11 mm; 11 to 20 mm; 21 to 50 mm; >50 mm), is shown for the multiple endpoints of the multinomial logistic regression models, which are (a) true interval cancers; (b) false negatives; (c) minimal-sign cancers and (d) occult tumors. The reference category (Ref) is screening-detected cancers. The black squares and the horizontal lines represent the odds ratios (OR) and corresponding 95% CI, respectively. ORs are presented on the log scale. Tumor phenotype = Luminal A: ER/HER2- or PR+/HER2-; Luminal B: ER+/HER2+ or PR+/HER2+; HER2: ER-/PR-/HER2+; Triple-negative: ER-/PR-/HER2-.