So the bottom line is that without a reasonable explanation for this approach, it's hokus-pokus?
There is a reasonable explanation for RDT. You have to make sure the drug candidate produces very high DCR, otherwise, too many patients have to discontinue before randomized phase, and among the DCR, most of them should come from SD, not CR or PR, otherwise, most patients left will be in open label again because only SD patients go on to be randomized. Overall, RDT is a much more complicated trial, used only on appropriate drug candidates.