Two papers from an innovative research initiative called I-SPY 2 were published this week in the New England Journal of Medicine. These multicenter, phase 2 trials involved patients with high-risk primary breast cancer in the neoadjuvant (pre-surgery) setting. Instead of using a fixed framework of assumptions to determine the trial’s parameters (as in traditional trials), these trials react to results as they arrive, a process called adaptive randomization. Patients with different disease subtypes were assigned to various experimental regimens based on their probability of responding to that treatment. Adaptive randomization speeds up the early identification of therapies that work well for specific disease subtypes, and helps avoid exposing patients to treatments unlikely to benefit them. In patients with specific subtypes of breast cancer, the I-SPY 2 studies found the drug neratinib and the drug combination veliparib plus carboplatin were more effective in eradicating tumours before surgery than standard chemotherapy alone. These types of adaptive studies can help to identify which of several promising therapies should quickly move to confirmatory phase III trials – fast-tracking the drug development process.