Drug-development, dose-selection, rational combinations from bench-to-bedside: are there any lessons worth revisiting?

One of the most intractable problems in drug development is getting an impartial and accurate evaluation of the agent one is working on, especially since patients are involved in clinical trials. It is particularly acute in the oncology field, where both cause and appropriate treatment of a tumor may partially be rational guesswork. In this situation, patients have a high likelihood of a sub-optimal outcome, notably when they, as profoundly impaired "salvage patients", end up in early clinical trials, receiving modestly characterized agents, which can do great harm if not judiciously used. These factors argue for caution when evaluating an agent, and there are several rules of thumb designed to terminate weak drug candidates as rapidly as possible. This precautionary stance is justified by the high failure rates of drugs in clinical trials. In this review, we argue that certain rules of the drug developmental paradigm are frequently misapplied, resulting in many good agents being discarded, and we provide an example of Hsp90 inhibitors. These "rules" are worth revisiting so that new cancer drugs can be developed and tested in a way that increases the overall success rates in patients.