Background: Early stage ER+ breast cancer may be treated with hormone therapy and/or chemotherapy. Currently available prognostic tools (e.g. Oncotype DX, Mammaprint and ProSigna) assess the risk of recurrence and the benefit of hormone and chemotherapy. However, these tests do not outperform traditional parameters such as tumor size, grade and patient age. New assays are needed to improve the prediction of relapse in ER+ patients. Methods: A continuous risk score (INDUCT score) predictive for relapse was developed using integrative analysis of publicly available microarray datasets and FFPE- based qRT-PCR assay. The INDUCT score was validated in (1) Affymetrix datasets consisting of 5 independent patient cohorts and 2) in a cohort of METABRIC dataset. The INDUCT score was further compared to traditional clinico-pathological parameters as well as to current prognostic assays such as Oncotype DX, Mammaprint and Prosigna. Results: The continuous INDUCT score (0-100) was developed to predict relapse in patients with ER+, lymph-node negative (LN-) breast cancer. A low INDUCT score (<42) based the expression of five genes (ESPL1, MKI67, SPAG5, PLK1 and PGR) identified a good prognosis group (8 year relapse-free survival probabilities ≥ 0.93) of 68% patients in both Affymetrix and METABRIC datasets. Importantly, the high expression of more than two genes was needed to achieve a score greater than 42, suggesting that the combinatorial effect of these genes is required to promote metastasis. The INDUCT score is significant in both LN negative [P=1.7x10^-4; HR=3.4 (95%CI 1.7-6.9)] and positive [P=2.8x10^-5 ; HR=2.5 (95%CI 1.6-4.0)] patients and in the presence [P=1.9x10^-5 ; HR=2.7 (95%CI 1.8-4.1)] or absence [P=5.6x10^-4; HR=4.0 (95%CI 1.8-9.3)] of tamoxifen therapy INDUCT is not improved by the addition of clinico-pathological parameters in LN- breast cancer and exhibits superior prognostic and predictive ability than Oncotype DX, Mammaprint and ProSigna assays. Conclusions: The five-gene INDUCT score demonstrates superiority to the traditional parameters as well as current prognostic assays in predicting the likelihood of relapse. This may further improve personalized management for patients with ER+ breast cancer.