I love the idea to do a "Master Protocol" for lung cancer (and eventually all cancer) registration trials. This is one good solution to the problem of opening trials for patients with rare mutations who otherwise might never reach a center that has a trial for their type of cancer, and would allow smaller centers to open the trial without breaking the bank on a trial that will only enroll one patient per year.

I do question the choice of using overall survival as an endpoint. Randomized trials of very effective targeted treatments (such as EGFR inhibitors in EGFR mutation+ NSCLC pts) typically do not show OS benefits as patients can get access to the drugs after progression on the control arms. In a Master Protocol, it is hard to imagine patients randomized to the SOC arms not being able to access effective therapies at the time of progression if the drugs being tested are truly effective (as the ROS1 presentation at ASCO this year taught us we can tell when a drug really works with only 14 pts). For example, I doubt the crizotinib randomized trials will ever show an OS benefit over chemotherapy now that crizotinib is widely available. Progression-free survival would make more sense for such a trial.