Replies to post #169252 on NorthWest Biotherapeutics Inc (NWBO)

[flipper44]

I get it Dan. Basically, what you are saying is that AVII is arguing two things at once that are 100% contradictory if one understands the immunotherapeutic interpretation of psPD.

If he is saying there was EARLY (aka: first interim) PFS futility due to psPD, then he is saying DCVax-L is not grapefruit juice and has a very high immunotherapeutic effect. However, at the same time, he would like to argue this would not then result in higher OS, which is contradictory to what is known about psPD.

(His garbage in garbage out logic then adds a cherry on top that he only thinks the blended is at 20 OS. This contradicts Dr. Bosch's August 31, 2017 presentation reference referencing 22 OS or how much higher can it go?) (AVII's current conclusions also contradict something AVII said a few years back referencing the IMUC trial -- aka: modeling does not work.) (Note: Not certain any of AVII's conclusions on this are relevant, because it may be DCVax-L has a different MOA to achieve longevity in non-psPD -- namely using the immune system to attack microscopic transitory mesenchymal cells.)

What I do think AVII has made a case for is that PFS may be harder to pin down, and could cause some false negatives. IMUC believes this happened in their trial to some degree even though their phase II was statistically significant for PFS. Huh? Big old Roche gets their questionable Avastin approved on secondary PFS with tons of side effects, increases death in those under 50 (with no black box warning on this), but IMUC, with little to no side effects, gets skunked by regulators, even though it also had successful secondary PFS and better trending OS. Shall we have the SEC investigate that?

Anybody think the FDA was equitable? Anybody?

[AVII77]

Dan, Thank you for taking the time to further explain your view. I understand what you are saying.

First, I agree with most everything you said. There is one crucial point however that explains why you see a contradiction where I do not. You said:

So having a futile PFS recommendation, if it were true, could only mean one thing: mPFS for the treatment arm were much worse than that of the control arm, .....

That is not correct.

If it were correct, I would not only agree that I was contradicting myself, I would agree NWBO is a promising investment.

This was discussed some time ago and I prepared a figure illustrating the futility scenario (at the bottom of this post).

That graphic was modified from the SPRINT trial and shows interim results of that trial (it was not an industry trial, or registrational trial, hence many (5) interims. They stopped after two consecutive interim analysis crossed the upper stopping boundary for overwhelming efficacy).

On that graphic, I roughly sketched a futility boundary in blue.

The bottom x-axis is "information fraction", a 60% interim is at 60% "information fraction" or 60% of the targeted number of events. You can see that for late interims the futility boundary rises above Z=0 (the Z score is the test statistic and can be translated into a p-value; a p=0.05 translates into a Z score of 1.96) and approaches the final p value (as an extreme example, if you do an analysis after 95% of events and see a p=0.07, the trial is unlikely to achieve p=0.05 with a few more events and thus it is "futile")

A Z score below 0 is trending towards harm and a Z score above 0 is trending towards efficacy. To stop early for overwhelming efficacy requires a higher Z score (a lower P value) than is required at the end of the trial.

Because of the alpha spend at the interims, the p value at the end of the trial has to be reduced (here, at the end, to claim SS at the end would have required p=0.045).

But the point is: at late interims, a trial can trend towards efficacy and still be futile. Any value below that blue line would indicate futility.

When I first shared this graphic I noted that: If I knew the interim PFS trended BELOW zero I would think this may be a good sign as it would suggest that DCVax does "something". Yes, it would still mean the Primary endpoint would fail, but it would give some hope that that "something" would translate into a survival benefit. And this is consistent with your observation and your interpretation of the contradiction.

Few here understand what futility means. And many argue that it isn't possible simply because they don't understand.

My earlier post on the topic is here

I truly hope this helps you better understand our difference of opinion.