Algorithm-assisted individualized therapy design improves survival in a mouse model of triple-negative breast cancer.

Chemotherapy remains indispensable in the treatment of malignant tumors but is often limited by the prevailing "one size fits all" approach, which neglects inter-patient variablity in pharmacokinetics and treatment response, often resulting in suboptimal outcomes. In this study, we explored individualized chemotherapy protocols in a clinically relevant mouse model of breast cancer using a novel algorithm-assisted therapy design (AATD). Two strategies were applied: a two-stage computational therapy protocol designed to stabilize blood concentrations of pegylated liposomal doxorubicin (PLD); and a model-predictive approach that optimizes dosing based on individual tumor characteristics. Compared to the standard maximum tolerated dose protocol, AATD-based personalized chemotherapy, guided by real-time monitoring of treatment response, tumor growth, and drug concentrations, significantly improved overall survival. Our findings in a mouse model of triple-negative breast cancer provide compelling evidence that chemotherapy can be personalized and optimized through algorithm-assisted therapy design.