Microarray technology and high-throughput proteomics have revolutionized cancer biology by generating vast amount of data for various cancers. To gain insights into the biological processes that drive breast cancer recurrence, we integrate gene expression profiles of breast cancer and protein interaction networks in search for affected regulatory and signaling networks statistically associated with endocrine resistance. To perform this integration systematically, we introduce a systems biology approach to screen related protein-protein interaction networks for active cliques, i.e., connected regions of the network that show significant changes in expression between recurrent and non-recurrent tumors. The experimental results from two breast cancer data sets show that the identified sub-networks can be effectively used to stratify patients treated with endocrine into groups with different outcomes. More importantly, the sub-networks can further help us gain mechanistic insights into estrogen action and endocrine resistance.