NorthWest Biotherapeutics Inc (NWBO)

[gpb]

On point (5), I think you're being overly optimistic regarding control versus treatment composition. I spent several days gaming out various scenarios of enrollment pace and PFS in control versus treatment based on all available data points, to see what might line up approximately with when the 66th event occurred as well as projected enrollment completion and topline and so on. This included 5 basic enrollment timelines (accounting for both the first pass 33 patients and the resumed enrollment in 2011 and sites that filled or withdrew and so on) and 7 pfs splits derived from the first 3 enrollment timelines (the other two enrollment schedules when extrapolated were too unrealistic and would have had us complete as of today or end-october respectively). The only way we're facing almost exclusively control-arm events is if the control arm did several months worse than historic SoC (optimistic enrollment pace), just plain flopped over a few days after being randomized (my best guess enrollment pace), or were counted multiple times (conservative enrollment pace - in other words, impossible).

That being said, even including a significant number of treatment arm events, we could still be talking about a treatment advantage of anywhere from 5 to 13 months (maybe a little more, but that requires plugging in some pretty optimistic variables - like going from aggressive-care back to historic-mean for control arm estimates). Of course, I absolutely cannot KNOW this, but it's my guess based on a hard review of exactly when every site was made active (or inactive), per-site enrollment rates adjust for inclusion criteria from various other trials, and so on. Some things had to be ASSUMED (danger zone) - for example, I assume that the treatment arm PFS is consistent between those in the initial 33 group's treatment arm and the subsequently enrolled treatment arm, and that no less than 26-33 of the initial 33 were in the treatment arm (company statements are unclear - "tended to drop out" - as to whether every single control patient from before the blinding withdrew, or just the majority of them, so I assume 33 and margin of error allows for 26). There are wildcard variables so I could be entirely wrong, but my point is, even as a hopeful long I do not believe we can be talking about "almost exclusively control arm events."