First of all, he never mentioned "majority" of over 300 patients. He said from contacts at sites, not "all sites". Some trial site investigators are feeling the sense that the whole batch is living longer than expected at this stage in the trial. Of course it's blinded, but when the same patients keep coming back for treatment and aren't dying it starts to become fairly obvious that something's occurring. Blinding means nothing when this is happening and it's pretty hard to keep that under wraps completely. That still doesn't make the experience translate into final analysis, but it sure starts to smell that way.