Prioritization of patients based on the risk of resistance to treatment plays a significant role in personalized therapeutic planning and improving disease course and outcomes. We have developed a series of computational algorithms to utilize patient molecular profiles to predict their favourable or poor response to therapy. We have demonstrated that a pathway-centric approach is superior compared to gene-level analysis alone. Further, we have shown that multi-level Big Data analysis to elucidate dysregulated pathways that govern therapeutic response is superior to using any single data type. We applied our approaches to chemotherapy resistance in lung adenocarcinoma and colorectal cancer, alongside tamoxifen resistance in breast cancer and have demonstrated their robustness, high accuracy, significant ability to predict treatment response and generalizability. Importantly, our predictions were not affected by common co-variates or markers of overall disease aggressiveness. We propose that the identified pathway markers can be utilized to prioritize patients who would benefit from specific treatments and patients at risk of resistance that should be offered alternative regimens.