Background: Breast Cancer Index (BCI) is a continuous risk index based on the combination of HOXB13:IL17BR (H:I) and the Molecular Grade Index (MGI) that estimates the individual risk of recurrence in ER+, LN- breast cancer patients. In the current study, a modified BCI model was developed using untreated breast cancer patients in order to evaluate its pure prognostic value, and to better optimize BCI for both early and late risk assessment. Methods: A model was built by linearly combining H:I and MGI weighted by their corresponding Cox regression coefficients using ER+ LN- patients from the untreated arm of the prospective Stockholm trial (N=283). Validation was performed in 2 independent ER+, LN- cohorts: the TAM arm of the Stockholm trial (N=317), and a multisite cohort of TAM-treated patients (N=358). Correlation of BCI with distant metastasis was evaluated by Kaplan-Meier analysis using the log rank test, and multivariate analysis adjusting for standard prognostic factors was performed using Cox proportional hazards. Results: The BCI linear model was significantly associated with risk of cumulative (0-10y), early (<5y) and late (≥5y) distant metastasis. Based on pre-specified cutpoints, BCI classified 64% and 55% patients as low-, 21% and 22% as intermediate-, and 16% and 23% as high-risk, with 10-y rates of distant recurrence (95% CI) of 4.8% (1.7-7.8%) and 6.6% (2.9–10.0%), 11.7% (3.1–19.5%) and 23.3% (12.3-33.0%), 21.1% (18.5–32.0%) and 35.8% (24.5–45.5%), in the Stockholm TAM and multisite cohort, respectively. Conclusions: BCI demonstrated significant prognostic performance beyond clinicopathological factors to predict cumulative, early and late risk of recurrence in early stage breast cancer patients. Use of BCI at diagnosis should enable clinicians to identify patients who are at high risk of late recurrence and may benefit from an additional 5y of hormonal therapy.