PURPOSE: Identifying biomarkers of radiotherapy (RT) response is important for optimizing the treatment of early breast cancer. In this study, we tested the interaction between endothelial cell (EC) expression of phospho-Tyr397-FAK (pY397-FAK) and adjuvant-RT on clinical outcomes after breast-conserving surgery (BCS) within a randomized study. Preclinical data suggest an enhanced effect of RT on low EC_pY397-FAK expression. EXPERIMENTAL DESIGN: We analyzed tissue microarrays from the Swedish Breast Cancer Group 91 Radiotherapy (stage I-II, lymph node-negative) breast cancer cohort, consisting of 1,178 patients randomly assigned to receive either BCS alone or BCS plus adjuvant-RT. Tissue microarray sections were immunostained for pY397-FAK, CD31, α-smooth muscle actin, and pan-cytokeratin. HALO analysis scored mean pY397-FAK intensity in CD31+ ECs, pan-cytokeratin-positive tumor epithelial cells, and α-smooth muscle actin + mural/stromal cells per core. For 822 patients, multivariable Cox regression analysis was performed for the primary and secondary 5-year endpoints, locoregional recurrence and all recurrence, respectively, as dependent variables and RT and EC_pY397-FAK as independent variables. RESULTS: EC_pY397-FAK expression was not predictive for the primary endpoint locoregional recurrence (P = 0.098), but the direction of the RT effect was in line with preclinical findings. For the secondary endpoint all recurrence, there was a significant interaction (P = 0.026) between EC_pY397-FAK and RT. Without RT, higher EC_pY397-FAK expression resulted in a lower risk for all recurrence (HR = 0.74 per SD; 95% confidence interval = 0.57-0.96; P = 0.026). CONCLUSIONS: Within the first 5 years following BCS, patients with low EC_pY397-FAK expression derive greater benefit from RT than patients with high EC_pY397-FAK expression. However, without RT, low EC_pY397-FAK expression is associated with a higher risk of recurrence.