Background: The 21-gene assay (ODX) has been endorsed by ASCO and routinely used in guiding clinical decisions on chemotherapy for estrogen receptor (ER) positive, node negative breast cancer patients. Many investigators have tried to identify more effective and economical methods that have predictive powers similar to the 21-gene assay. In 2013, Klein et al reported that the Estimated Recurrence Scores (ERS) from three Magee equations derived by linear regression analysis demonstrated relatively high concordance with the RS of ODX. The current study aims to evaluate the clinical and prognostic utility of the three Magee equations.<br>Design: 560 patients with ER positive tumors, clinical follow-up and diagnosed between 1997 and 2012 in our institution were included. Modified ERS from all three equations, which used the Nottingham grade (3-9), modified H-score of ER and progesterone receptor (PR) (average staining intensity X percent positive tumor cells, 0-300), HER2 status (negative, equivocal, positive), tumor size (cm) and % labeling for Ki-67, were calculated from each case. A mean modified ERS of each case was used for statistical analysis.<br>Results: For this cohort, the mean age was 60 years, the mean tumor size was 2.17 cm, 71% were histologic grade 1 and 2, 63% were node negative. All patients were ER positive, 85% were PR positive, and 10% were HER2 positive. The Modified ESR was not associated with patient age, tumor size, or histological grade; but it was associated with nuclear grade and expression of ER, PR, HER2 and Ki-67. Although nodal status was not included in calculations in these equations, a high Modified ERS (>30) was strongly associated with positive node. Among the three components of Nottingham grading, only nuclear grade showed a significant association with Modified ERS. Only 4% of ER 71-100%, 2% of PR 71-100%, 8% of HER2 negative, and 5% of Ki-67<15% had Modified ERS >30. Also, Modified ERS was significantly associated with progression free survival in these 560 patients (p-value=0.0031) and in node negative subgroup of these patients (p value=0.0069), but not in node positive patients (p value=0.8697).<br>Conclusions: Modified ERS can effectively stratify ER positive patients into different prognostic subgroups. It may be used to guide clinical decision making for patients with ER positive tumors, in conjunction with other clinical-pathologic factors.