The trial was not halted, it was only a partial halt. I don’t think the reason for the partial halt is as you suggest, that a full halt, for the reasons you suggest, would be a different event and would have had to have been announced, IMHO, as it would have been a major regulatory event. I do think the crossover FEATURE of the trial, which was set from the beginning because no one would enroll, was a known potential issue. Crossover trials have a feature that is rarely discussed here that every patient is their own placebo and treatment arm. So not sure that that is fatal either. Then lastly, the regs have been re-written in the various countries to allow the ECA. So while it has not exactly been done like this before, it is not necessarily problematic.
I think the reason for the partial halt is answered in the lack of a substantial number of placebo arm patients, not in treatment patients. If the placebo arm is no longer happening, then subjecting patients to a delay in treatment is unnecessarily damaging to them. If one is going to move to an ECA, it would be unethical. And given what they knew at that point and the advice they were likely being given from regulators about what was coming, these reforms take many years, I think they knew that continuing was the best route, and that taking in more placebo arm patients was basically no longer feasible and they would get the number of patients in their treatment arm that much faster when everyone already enrolled became a treatment arm patient. There is also the very strong likelihood that the regulator and DSMB agreed on the adjustment without unblinding them in any way, as the Federal Court judge indicated and the interim analysis indicated was the case. So the adjustment to ECA was basically necessary with each regulator, the regulators fully in the know. The adjustment to OS seems to have been a likely necessary and obvious change for much longer than they likely wanted to discuss because the changes would allow critics to make all sorts of unlikely claims, as the court case and Washington Post article discuss.
So they remained silent and followed the previous path until the regulators, lastly the FDA, made a change in rules for ECA and until they had accrued enough time for all patients enrolled to change to OS from PFS, meaning even the last patients enrolled, you want a good amount of time to pass to ensure you have accumulated the time reasonable to calculate OS for all patients. I think the regulator was well aware of what was going on and was basically leading them by the nose though of course the approach of the regulator in the public course of regulation is that they provide no advice and are not responsible. But the reality is likely much closer to the reality much of the time with any regulator, in the US, you consult regularly, and get a lot of direction from that regulator, through informal consultation that is entirely deniable because the consultations are entirely informal and not part of the main process that everyone knows about. It comes with informal calls, lawyers to regulator, question and answers about pending reforms and how best to proceed.
Market Data 
Markets 
AI
