Purpose. To determine intrinsic breast cancer subtypes represented within categories defined by quantitative hormone receptor (HR) and HER2 expression. Methods. We merged 1,557 cases from three randomized phase III trials into a single data set.These breast tumors were centrally reviewed in each trial for quantitative ER, PR, and HER2 expression by immunohistochemistry (IHC) stain and by reverse transcription-quantitative polymerase chain reaction (RT-qPCR), withintrinsicsubtypingbyresearch-based PAM50 RT-qPCR assay. Results. Among 283 HER2-negative tumors with ,1% HR expression by IHC, 207 (73%) were basal-like; other subtypes, particularly HER2-enriched (48, 17%), were present. Among the 1,298 HER2-negative tumors, borderline HR (1%-9% staining) was uncommon (n 5 39), and these tumors were heterogeneous: 17 (44%) luminal A/B, 12 (31%) HER2-enriched, and only 7 (18%) basal-like. Including them in the definition of triplenegative breast cancer significantly diminished enrichment for basal-like cancer (p , .05). Among 106 HER2-positive tumors with ,1% HR expression by IHC, the HER2-enriched subtype was the most frequent (87, 82%), whereas among 127 HER2-positive tumors with strong HR (.10%) expression, only 69 (54%) were HER2-enriched and 55 (43%) were luminal (39 luminal B, 16 luminal A). Quantitative HR expression by RT-qPCR gave similar results. Regardless of methodology, basal-like cases seldom expressed ER/ESR1 or PR/PGR and were associated with the lowest expression level of HER2/ERBB2 relative to other subtypes. Conclusion. Significant discordance remains between clinical assay-defined subsets and intrinsic subtype. For identifying basal-like breast cancer, the optimal HR IHC cut point was ,1%, matching the American Society of Clinical Oncology and College of American Pathologists guidelines. Tumors with borderline HR staining are molecularly diverse and may require additional assays to clarify underlying biology. The Oncologist 2015;20:474-482Implications for Practice: This study pooled centrally reviewed hormone receptor (HR) and HER2 data and individual gene expression and intrinsic subtyping from three cooperative group trials.The results indicated that the optimal cut point for defining triple-negative breast cancer, if the goal is to enrich for basal-like biology, is to adopt the guideline of ,1% staining. Tumors with borderline HR expression are highly biologically heterogeneous, which raises the question of whether these tumors should be considered indeterminate. A proportion of clinically defined HER2-negative tumors were defined as molecular HER2-enriched subtype; however, whether they are suitable for anti-HER2 therapy needs to be determined.