Celebration Thread! Biogen is going to the FDA with Aducan.

Alzheimer's, cardiovascular, and other chronic diseases; biomarkers, lifestyle, supplements, drugs, and health care.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

Correlation Hold Back 2.gif
You do not have the required permissions to view the files attached to this post.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

Above orange line is a free hand of the regression for the 302 study (using the 103 proportions).
The 301 low dose dot would then fall closely into line beside the 302 low dose.
From above the 103 1 mg/kg was considered too high and the titration and 10 mg/kg too low.

If all of this has not actually already been adjusted for, then the alignments of these true points could be almost magically good.
This would mean that the 103, the 301 low dose and the 302 low and high dose arms all form one explainable dataset. This involves hundreds and hundreds of patients. The origin point would also represent a very large number of patients. The Briefing Document never tried to integrate all of the patients into one framework. Doing this helps to amplify the statistical power of all of the patients and creates a seemingly even more highly persuasive way of understanding the results than that was presented in the document. This style of thinking dramatically reduces the plausibility that the trials were some sort of a fluke. All of these points (except the 301 high dose)
now apparently can reasonably be placed along a single regression line. Further we can think of the 302 high dose as a large hold sample that apparently fits the expected adjusted regression remarkably well.

This feels as if all the points are moving into focus and we can speak of a single law of amyloid anti-body cognitive response.
Basically this law says that a given demographic of MCI/Mild AD patients (46:54) (though this proportion could be chosen somewhat arbitrarily: it would however be helpful if future trials actually accepted some cohort adjusted standard AD population for the purposes of comparison 50:50?) will exhibit a linear response along a slope of ~2.88 with relation to amyloid reduction measured in SUVR. Once this can be more fully clarified than other treatments might simply only be required to established amyloid reduction (which is easy) and the relationship with cognition will then be considered a given.

Next up might be to look at some other AD clinical trials and see how they fall on the Alzheimer regression.


One additional idea that I want to include is that the differences between the different demographic is truly large. The Mild AD patients had a 2.88 placebo decline versus 1.53 for the MCI! This is very large. This might explain why we saw shifts in the trial after PV4. Simply moving to more Mild AD patients would lead to the seemingly unlikely leaps that were shown in the Documents. Moving to more milds would do that. It also speaks to the large treatment effects that were replicated in 302 high dose. When you realize that they used an 80:20 mix (302) instead of a 46:54 (103) you can understand how the seemingly large discrepancy could arise. It was merely because they included more slow decliners in 302. It is hard to believe that this highly deceptive presentation of the figure would be included in the Briefing Document (perhaps this was in fact corrected for. Though to the best of my ability it does not appear that it was). When you make some of these corrections all of a sudden the points really feel as though they are almost right on the regression line. Of course, you do not want to over correct. Probably best to only make one correction for each dot because everything is highly linked into different correlated variables. For example, the men: women difference probably relates to the dosing difference noted above, though this might also be expressed in differences for nations, possibly apoe4 genotype etc.. Trying to account for only one difference might be the best.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

So, it is remarkably clear where I am heading with this. I can now keep on throwing darts at the board and prespecify (with some degree of accuracy --possibly hit a fair number of bulls eyes) and at some point I mean an honest skeptic would have to give up. Perhaps, even to make this interesting I could ask around for those might want to forfeit their money on a bet. These points can all largely be seen to be collinear after simple adjustments are made (especially standardizing the population mix to the 103 trial). There are already about 10 points that seem to fit quite well to the regression line. A single line that defines MILD/MCI AD treatment efficacy with amyloid anti-antibodies; though it is true that perhaps some emerging antibodies might be of a different class so that the dosing relationship might no longer hold.

The basic idea is that you can take a nice big point like the 302 high dose (or others) and just situate it just so. With all of the hundreds of patients, such a point has a great deal of leverage and it also has a large amount of center of mass. Such a point would almost have to be at its "true" position. Then all of the other points would be expected to fall into place. The 302 and the phase 2b Ban2401 serve as fairly solid anchor points. The only question that arises is finding the weightings that were used for MILD and MCI and their effect sizes. This is not always easy to find. We might wind up somewhere up to a slope of ~1/0.278 = 3.6. Perhaps in the range of (3.4 - 3.6). What had been somewhat hidden in the document was actually how large the MILD CDRsb effect was. Around 1?
That is not exactly a small effect. It was only by weighting it 80:20 MCI:MILd that this became obscured. It is very unclear to me why this was done. Highly weighting those patients who are not likely to decline much makes showing a benefit more difficult. The average decline of the MILDs was 2.88 which is big. Basically time has selected those patients who truly do have clinical decline. Younger patients might more be motivated to join the trial because of what they fear might be. If the trial had been weighted 20:80, then it would have been an easy success. The MCIs could then go off label.


All those points all on the same line within a integrated framework greatly narrows the plausibility that they somehow all managed to fall so nicely on the line. Even if someone were to argue against every single dot (individually) on the figure as being valid, and perhaps I would be willing to give concede every single dot (individually) yet the clear linear trend remains. The collective power of the totality of the dataset is overwhelming. I am not clear why such an approach was not presented in the Briefing Document.

Of course, the big thorn in the side of this discussion is the 301 high dose dot that appears to mock all of others from its own celestial strata. How does the dot belong? The Briefing Document describes the issues involved. One of the more prominent problems was the outliers. It is remarkable that some of the patients had CDRsb declines of upwards of 10 points! That is truly startling. How could that have possibly been allowed? The entire clinical trial integrity becomes open to challenge. Each 10 pointer has the weight of 25 0.40 pointers. This is the tyranny of the outlier. In some of the Figures in the Briefing Document some of the anomalies of the 301 study simply disappear when these outliers are removed. It is the old there are 10 people with $1 and 1 person with a 100 million. Think about how wealthy everyone is on average. Um, how about at the median? There should be pre-specified fall backs to retreat to when the average does not reflect "the average". When you actually look at the 301 patients with high dose exposure, they perform almost exactly the same as those in the 302 study. The MILD AD high dosers alone could easily contribute -0.42+ to the 301 high dose dot (For the 302 comparison, MILD Ad high dose had a treatment effect of -0.95 *0.46). It is quite easy to imagine that the 301 high dose could realistically be transferred from the clouds to its place on the line. However, this is supposedly an unmentionable truth.

Other dots could also be added. I looked up the BAN2401 phase 2b results, it appears another ~900 patients worth of regression line confirmation are available. The top dose hit ~0.30 SUVR and 26% CDRsb improvement. So that would place the dot pretty much exactly on the regression line. The 302 high placed at -0.80 seems a little low now, 26/22 *0.80 =0.95. That is exactly where the original regression equation placed ~ -0.27. Slope m= ~3.42 (from 2.96). There are other points to place from this phase 2b trial.
I think I included this early in the thread, though I will add in the additional points to the figure.

There is now also another anti-amyloid donanemab reported a 32% decline on iADCS. This one might also fall on the line. When is it time to just call it quits? All of this points following a simple formula are moving reasonable discussion further and further away from plausibility.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

anatomy 1.jpg
You do not have the required permissions to view the files attached to this post.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

Sorry everyone I don't want to bother you with my problems I am sure that you have enough of your own, though I have not of late been feeling 100%. I have had a somewhat mysterious pain on my left side that I have been trying to understand. The pain appears to be located at the red dot in the above figure (in the red rectangle, level with my belly button). When I looked up what this part of the anatomy actually does I found that it does not appear to do that much at all. Basically, I have a mystery pain in a body part that has no clear functional role. No major functional organs ( aside from the digestive system): Sardonic Medicine.
Wonderful.

Fortunately, I seem to be making progress (ergo this post; if I were stomped, I would probably have kept this to myself). What it seems to be is something in the hernia type class of medical ailments. I have actually carefully felt around and there does not seem to a true hernia so it might be more of a muscle strain. How could such a misfortune arise? Lately a family member has become a viewer of podosphere media and about every 5 minutes for an hour every day will race to my room to report the latest outrage (or at least whatever the latest 2 minutes of hate of the last 5 minutes has reported. Yes, online media clearly is moving us to this hyperdrive emotional state without rational executive control. A grown level conversation about how we want media to shape our reality would seem overdue.) Every time I hear the approaching gallop of foot steps I stand up from my very comfortable low rise chair. It is basically like doing a quarter squat repeatedly, I typically lean to the left when I do this squat. That is my best guess of what has caused my problem. Online podcasts have caused my abdominal muscles to weaken. Given the limited resources in today's medical system I will probably need to think up a more presentable cover story than the truth, if this needs to go mainstream medicine.
Rejecting this medical problem because it is simply too dumb (Sorry J11 we just don't have the budget for it.) would add considerable insult to injury.


In the future I will remain seated in my highly cosy and comfortable chair and welcome my family member in without any further exacerbation to my lateral hernia/ muscle strain. This already seems to be helping. I also noticed that drinking water helps; while eating makes it worse. Hopefully I will soon be back in top form and can return to my high frequency posting on this thread: Unless there is some weird medical problem that can arise from straining some part of the anatomy in such a pursuit.
Last edited by J11 on Thu Feb 25, 2021 9:24 pm, edited 2 times in total.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

BAN2401-study.jpg
ban2401-ererer.jpeg
14.jpg
You do not have the required permissions to view the files attached to this post.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

images.jpg
Profiles-of-Aducanumab-BAN2401-ALZ-801-14JAN21.png
You do not have the required permissions to view the files attached to this post.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

OK I wanted to have all these figures lined up for when I am feeling 100% and can create a single Figure for all the points. As we can see above it appears that the phase 2b BAN2401 trial is highly similar to the 302 aducan result Remarkably (almost magically)
1.5 * 0.26 = 0.39

My feeling is that when you perfectly nail the answer in a scientific experiment (chemistry, physics etc.) probably a good idea to add some noise in: Who will believe if it is 100% correct? 0.39 was the exact CDRsb result from the 302 high dose aducan arm. Forutnuately the amyloid SUVR was slightly more in the Ban2401 so it is not nominally exactly in line with expectation: Still clearly this result captures one's attention.

The phase 2b had significant numbers in the different arms : "by the end of enrollment, 161 people were on 10 mg/kg biweekly, 253 people on 10 mg/monthly, and 247 people were on placebo. Only 52 people were in the 2.5 mg/kg biweekly dosing group, 51 in the 5 mg/kg monthly, and 91 people in the 5 mg/kg biweekly groups. Following the EMA request, the second-highest dose group ended up chock-full of APOE4 carriers, at 89 percent, compared with 30 percent in the highest dose group, and 70 percent in the placebo group."
https://www.alzforum.org/news/conferenc ... ve-decline

There was a complication with assignment of APOE4s to the highest treatment arm mid-way through the trial.

I can add in all of these numbers into my figure.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

Perhaps we will also be able to add in some points for dona which has recently reported back positive. In the above figure with reported ~30% cognitive benefit would plausibly place it also on the regression line. The problem that we could face here is that they might report the treatment arm proportions that are MILD AD/MCI. This would make things difficult. The BAN2401 result of 1.5 decline on CDRsb is right in line with the 301 placebo.

I am not sure, at some point those arguing this will have to give up. I mean, ole J11 is going to keep on posting the numbers and at some point there might be universal agreement that all this overwhelming evidence is too much information. I will definitely consider such pleas as they arise.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Celebration Thread! Biogen is going to the FDA with Aducan.

Post by J11 »

"hominem unius libri timeo." (in the original meaning). I think that about covers it.
Post Reply