Methods: Patients treated in the aTTom trial with available primary tumor tissue were eligible for this multi-institutional prospective-retrospective study. Primary and secondary endpoints were recurrence-free interval (RFI) and disease-free interval (DFI), respectively. Statistical significance level for RFI was set at 0.0336 as per statistical plan. Kaplan-Meier and Cox proportional hazards regression analysis with time-varying coefficients were used to test the predictive activity of BCI by HoxB13/IL17BR (H/I) status (High vs Low). Likelihood ratio test based on Cox regression was used to evaluate treatment by biomarker interaction. Results: 2637 tumors were centrally assessed for ER, PR and HER2 status leading to 1822 HR+ patients analyzed (1018 N0, 583 N+). Initial results from patients with N+ disease at 12 years of median follow-up showed 287 (49%) were classified as H/I-High and 296 (51%) were classified as H/I-Low. H/I High patients showed a statistically significant benefit of 9.8% in RFI with 10y vs 5y of TAM (HR=0.35 [95% CI 0.15-0.85]; P=0.027), whereas H/I Low patients showed no benefit (-0.2% RFI; HR=1.07 [95% CI 0.69-1.65]; P=0.77). A statistically significant interaction between continuous BCI and treatment was demonstrated (P = 0.02). Conclusions: These data provide further validation and establish level 1B evidence for BCI as a predictive biomarker for preferential benefit from EET in HR+ breast cancer.