Abstract Abstract #1068 Aim
 Identification at the molecular level of breast cancers sub-types associated with different clinical outcomes would be of great value to help individualize therapeutic strategies and, in turn, improve survival. With large sample size, long follow-up, and geographical spread of the population any results derived from analysis of the Nurses' Health Study (NHS) data set may be more generalizable to the U.S population. Thus the purpose of this study was to define the survival outcomes associated with distinct molecular phenotypes of invasive breast cancer in women identified from the NHS.
 Methods
 2013 women enrolled in the NHS (1976-1996) with invasive non-metastatic breast cancer whose breast tumor samples were available for inclusion in tissue microarrays and subsequent immunohistochemical (IHC) analysis form the study population. Tumors were classified into one of 5 categories based on results of IHC assays for estrogen receptor (ER), progesterone receptor (PR), HER2, cytokeratin (CK) 5/6, and epidermal growth factor receptor (EGFR) as follows: 1) Luminal-A (ER and/or PR +ve and HER2 -ve), 2) Luminal-B (ER and/or PR +ve and HER2 +ve), 3) HER2 subtype (HER2 +ve with both ER and PR -ve), 4) Basal-like (-ve for ER, PR and HER2 and +ve for either CK5/6 and/or EGFR), 5) unclassifiable (-ve for all markers). Overall survival (OS), breast-cancer-specific survival (BCS) and recurrence-free survival (RFS) were estimated using the Kaplan-Meir product limit method and compared across groups using the log rank statistic. Cox-proportional hazards models were fitted to determine the association of molecular phenotype with survival outcomes after adjusting for age at diagnosis, stage, lymph nodes, tumor size, grade and body mass index at diagnosis.
 Results
 Median age at diagnosis was 57 years (34 – 75 years) with a median follow-up of 14 years. 1490 (74%) tumors were classified as luminal-A, 99 (4.9%) were classified as luminal-B, 106 (5.27%) were of HER2 subtype, 219 (10.9%) were classified as basal-like and 99 (4.9%) tumors were unclassifiable.726 (36%) patients had died of any cause, 433 (21.5%) had died of a breast cancer related event, and 459 (22.8%) experienced a recurrence. Five-year BCS for patients with luminal-A, luminal-B, HER2 , basal-like and unclassifiable tumors was 94%, 82%, 73%, 82% and 75% respectively (p<0.001). In the fully adjusted multivariable model compared to patients with luminal-A tumors patients with luminal-B (HR 1.73, 95% 1.18-2.53), HER2 (HR 1.39, 95% CI 0.97-1.20), basal-like (HR 1.50, 95% 1.32-1.20) and unclassifiable (HR 1.89, 95% CI 1.30-2.74) tumors had lower BCS. Similar trends were observed for OS and RFS.
 Conclusions
 Compared to women who have luminal-A tumors those with luminal-B, HER2 subtype, basal-like and unclassifiable tumors had a worse prognosis. A fifth, unclassifiable sub-group was identified that has survival outcomes similar to and may represent a subtype of basal-like tumors. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 1068.