NorthWest Biotherapeutics Inc (NWBO)

[CDRD]

You wrote:

"Seems like there is still a lot of confusion on the board regarding median OS."

Yes there is.

You wrote:
"Let's start with the basics. When the trial reads out, is the reported mOS based on the group of actual patient events (say median of the 233 recorded events if unblinded then), or is it the estimated median survival of the entire trial group of 331 (so the 165/166 rankings) with the KM technique applied to account for events that have not actually occurred or been recorded (LTFU, etc.,)?

I am not sure how they account for the living patients but if the hypothetical you state below is carried out, the the MEDIAN WILL IMPROVE (Example given below).

You wrote:

"If it is the latter, and the remaining patients alive have all exceeded the survival time of the current 165/166 events in rank order, then shouldn't the median remain static going forward as additional events would be plotted to the right of 166 on the curve all the way out to 331?"

NO. The median will not remain static if the remaining patients are included and if there survival times exceed those of the deceased. Here is a simple example:

say that there are 5 deceased patients with survival times:

1 month
2 months
3 months
4 months
5 months

median is therefore 3 months

If a living patient (whose longevity already exceeds the 3 month mark)-is ADDED to the pile, that would place them in the rank order somehwere after the 3 months (let say the living patient was alive for 4.5 months at the time of the trial unblinding); Then,

1
2
3
4
4.5
5

would be the new rank order and the median is now between 3 and 4 (or precisely 3.5).

Go read eagle8 post #131659--he summed it well.