If every patient was able to receive chemo after chemo-radiation combo then that would be a good comparison. However, this was true of only 28 of the 454 patient that they looked at in that analysis. Those that received adjuvant chemo-radiation only after resection (246 of 454) had an OS of 25.2 months, which could be similar to the IMPRESS trial.
So I guess you have an explanation for the control group living for 28 to 30 months?
I still believe the vaccine works. I'm disappointed that they didn't use an alternative analysis to show a benefit if it is indeed there. Log rank wouldn't show a benefit of 30% improvement if the control group had a MOS of greater than 23 months, based on the censoring of patients, which we now can say is the case. The blended MOS for the trial is around 30 months based on my calculations so there's no reasonable explanation that the control group would live this long. I believe that the control group may have a MOS of 24-25 months with the vaccine arm around 36 months. Again, even with this improvement log rank wouldn't show a 30% improvement due to right censoring of patients that were enrolled less than 2 years ago.
It's possible that the vaccine doesn't work but I don't think that you can conclude this by the lack of a stoppage at this interim. If they were to use another statistical method and the trial would have continued then I would be less optimistic. I guess we'll all find out around the middle of next year.
I understand what you're saying, it just doesn't make sense. You like to ignore the fact that the high dose group had worse baseline characteristics because it doesn't fit your thesis of no dose response. That's totally up to you if that's what you want to believe. It's too bad that you don't realize that this is important along with the fact that you haven't even paid attention to the control group in this trial until I pointed out that this is also important.
"If you want to trade insults I'm not going to do it other than to say you are an uninformed clown."
It's good that you're sticking to facts instead of using insults.
Actually, making stuff up isn't the same as doing research. I'm glad that you found that piece and shared it, it was appreciated. But I don't claim to know things that I don't, unlike you. I'm not ashamed on making predictions but I'll say that's what they are, and I'll give estimates on how likely they are to happen. You just like to hear the sound of your own voice even though most of the things you say make no sense whatsoever.
I think the real question becomes, how likely is it that NewLink uses alternative analysis? In my opinion, if they use log-rank they aren't likely to succeed at this interim (10-20% success). If they use an alternative approach, they aren't likely to fail ( 95% success). So it all boils down to which approach they use. Hopefully their statistical consultants were able to communicate this dichotomy of possibilities and the FDA agreed to alter the SPA. I believe the longer it takes for the DSMC to deliver the results, the better.
If the trial continues, should trade around $35. If the trial is stopped, it depends on two different scenarios:
1. Log-rank analysis is used (roughly 30% improvement shown) - $120
2. Reduction of risk analysis is used (roughly 100% improvement shown) - $225
I don't think the 30% improvement requirement will change if they use the alternative analysis. 30% is tougher to prove initially but gets much easier to prove as time moves on. On the last CC NewLink said that they've spoke with statistical consultants and I believe that they know it would be really difficult for the trial to continue at this point if they change to alternative methods.
I've come out with about 22 months but look at Mind Craft's research as it's much more thorough than mine. His Seeking Alpha article is called "NewLink Genetics: Examining Likelihood Of Success In The Impress Trial". It really dives into the differences in survival based on patient characteristics that is easier to handicap in pancreatic cancer vs. other cancers.
I've mentioned this three times now, this will be the last. Look at the baseline characteristics for the two doses in the Phase 2 trial. The high dose group was much worse off compared to the low dose group. There's no way the high dose group should have lived anywhere close to 20 months let alone above it!
There have been many here that have sliced and diced the Phase 2 results over the past several years. At this point, I believe the consensus among the bulls at least is that you can make whatever assumptions you want out of that data but there are a few big differences in the Phase 3 trial that make understanding every detail about the Phase 2 a moot point, including more dosing, high dose for everyone, and healthier patients.
But even if you know absolutely nothing about the Phase 2 trial you could still find reason to believe that the Phase 3 will be successful for two reasons: the average OS for the entire IMPRESS trial is north of 30 months (based on enrollment figures and the two interim dates) and pancreatic cancer patients with the same baseline characteristics as those in the IMPRESS trial live around 20-22 months based on historic trials.
You could reasonable believe that the control group may live for a few months longer than this based on newer surgical techniques but more than 24 months OS seems HIGHLY unlikely. If the control group lives for no more than 24 months, the odds of success in the IMPRESS trial is well north of 80% in my opinion.
It's such a small number that it's difficult to extrapolate this to a large number of patients for the IMPRESS trial. Whether the MOS was 22 or 25 doesn't matter that much in the phase 2 for that reason plus the fact that patient's were in worse shape in the high dose group, like I mentioned earlier.
It only shows 15 of the 26 patients in the high dose group on that particular chart. I didn't see anything groundbreaking in the document, but still an interesting breakdown of some of the data. The abstract says "Figure 9 shows that the correlation of increased anti-a-Gal antibody production in patients with better outcomes is observed only in the group receiving high doses (300 million cells) of Algenpantucel-L.
This is interesting information that I haven't seen before, thanks. I'll look at it today and try to make some sense out of it.