Like what they said they would do for the PFS, they would wait until the OS for the whole trial stops moving out. As of this thursday, it will be 196 days since the last patient got on study. If you look at the shape of the control survival curve and sum the number of patients that would be alive from the control arm, you get a number in the 6 to 8 range. We are now well into the tail of the control group. So only a small fraction of the 25 or so control patients that were enrolled between July 15 and Sept 10 in 2012 are still around.
If you look at the shape of the PII REO curve, and sum the patients that would still be expected to be around, you get a number around 36-40. This arises from the size of the tail even at 400 days ..... quite a few (about 40 %) of the quartile that got on study in the 3 months preceding last March 31 will still be around at this point. The picture is of course better for the second half of the trial that got on study after March 31 2012.
At this point, the control OS is well established, but most of the 36-40 test patients will out live that value, pulling the obsevred OS for the trial out further. So when can we expect an outer limit?? It will occur at or near the time of the OS of the test group as more death events won't continue to pull the overall OS out. If you use the PII OS this will occur on about April 21. But, keep in mind that 1) patients on the REO arm will continue on treament with REO alone if they have not progressed (this wasn't done in the PII), 2) the PII had 38% 3rd through 5th line patients but the current trial will be almost all 2nd line, 3) Mets are being tracked now so if the mets continue to respond and the primary doesn't they still continue on treatment.
I think this won't unblind until late may/early june.
Nome: Some nice reasoning Nome; thank you. I like your answer better than mine. The company has chosen not to unblind the PFS and pay the statistical penalty, because of that I speculate that they may be shooting for some form of early conditional approval in mets, neo-adjuvant, local or some combination. The giggles from Matt and Brad would indicate that they are pretty pleased with themselves about something?
I do believe that they will be releasing results from some of their multiple studies just prior to the AGM; the optics almost demand it. What do you think?
Neber: I beleive that Jmodified is correct. To be certain that both groups have stabilizied survival times would require that 75% of the patients die. If the trial is experiencing long tails and if Reo is doing very well, then that could be a long period of time. My conjecture is that the trial will be unblinded based on a combination of events (deaths) and time. The company has likely researched every piece of evidence they can find to estimate expected survival times in the control group; they also have the unreleased data from their previous H&N trials. Therefore the company likely has a final timeframe in which they will unblind the results if patients continue to survive. I have no idea how long they will wait but my wild guess based on absolutely nothing, is an unblinding in the fall 2013.