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 »

Now that I have data massaged the aducan, ganta and Ban results what are some comments that can provide some context? What do all of those regressions mean?

Clearly, these 3 anti-amyloids in particular appear to closely follow a linear pattern over the indicated range of SUVR. While it is probably unrealistic to firmly believe that the points have 99%+ correlation (they were after data massaged), the true empirical relationship is still quite impressive. It is important to keep in mind that with clinical science the 100% correct result is almost by definition incorrect. Life has a certain definite amount of builtin wiggle to it. Correcting for this attenuation of reliability might produce a truly remarkable result. I am not sure how much that would apply to SUVR, though CDR-sb likely has such a correction that could be made and would put into focus the actual correlation. Yet, even without correcting for this or Censored data ITT (301 and 302 low) or even without switching to opportunity to complete for 302 hi, the relationship is strong. Why?

Why or more importantly how could such a result have occurred? I (mark the bullseye after throwing the dart) think it relates to the
impressive Alzheimer infrastructure that has been developed over the last decade or two. In order for the points to line up as they did for different trials using different anti-amyloids in different Memory clinics, highly accurate amyloid imaging was needed (SUVR), highly standardized patient categorizing was needed, a highly developed and standardized psychometric infrastructure was needed, among others. The Alzheimer's network that has been organized must have cost quite a bit of money and has required a fair amount of effort at a large scale. Improvements such as standing patient registries clearly would be helpful, though the existing network has developed the critical mass to produce the impressive results that I have documented in previous posts.

From what I can tell, aducan, ganta and Ban all follow the same regression. My take on this is that perhaps removing amyloid anywhere except at the monomer stage will cause the other amyloid species to equilibrate. I am not as clear about dona. It is possible that it is also on the regression. I wanted to remove it from the others because it had a lot of leverage being so far off to the left and I did not want it dominating the results for the others.

dona raises several issues.

1. For example, in an earlier post I estimated that maximal amyloid clearance would roughly occur at ~ -0.36; the dona results gave this as -0.372. It would be interesting to see whether going "below zero" might be helpful in AD. What happens is that AD patients arrive at the clinic with around 1.40 SUVR and then SUVR is brought down to ~1 and they are then described as amyloid negative and treatment can stop (this is what happened with dona). I am not clear whether they have went through the research on this question. Perhaps going closer to absolute 0 SUVR could be helpful for AD patients? Ina AD patients amyloid is not merely a nuisance brain protein (as it might be in healthy brains) but a focus of neurodegenerative illness. Perhaps more amyloid removal in the context of this ongoing neurodegeneration might be clinically helpful; unclear whether this is actually true.

2. Possibly the regression line is only valid within the range that was used in the posted regressions. After ~0.30 SUVR perhaps the regression line goes horizontal, more amyloid removal has no marginal benefit. This also is unclear.

3. What I found interesting was that the CDR-sb for dona was quite a bit different than the other amyloid treatments. aducan had CDr-sbs of ~2.4-2.5, similar result for ganta (2.0 -2.1), while dona had 3.4 - 3.6. Everything probably lined up so impressively for aducan, ganta and ban because they were from the same era of AD clinical research. It is possible that dona did not perfectly line up with these because it is now in a somewhat different clinical era; using tau selection and more advanced patients as measured by CDR-sb. It really is reassuring to have such good agreement with the three anti-amyloids, though for things to move forward new and even more powerful clinical research techniques (such as those used with dona) need to be allow to evolve. It is especially to remember that what can be described as truly pathological science has occurred in the past when scientists tried to chase after the right answer only to find that the "right answer" was wrong! It is important to create the space needed for innovation to occur.

4. The dona tau selection is an especially good example of this. Narrowing in on the actual responders will greatly increase statistical significance and the reported clinical benefit. Clarifying the actual responders would seem to be aligned with both corporate and patient interests. The great mass of the AD treatment population is actually before symptomatic illness. There might be 100 million in this worried well condition. If given a product many would be interested in early treatment. In the actual clinical research they are only thinking about those who can demonstrate a measurable cognitive/ functional improvement over the duration of a clinical trial.
The corporate worry is that all of these patients will be lost if they can be shown not to benefit. This really is not true. All they need to do is make a one time declaration that those with too much tau will not benefit from treatment. From then on all AD patients become eligible for their treatment. All future patients will move through the optimal tau treatment window and should derive substantial treatment benefit. In fact, treating early should be just as helpful. The only difference is that the benefit to treatment will be implicit not explicit. Implicit is good. Implicit means that patients will not have to experience the neurodegeneration at all; they will have been treated before it has even become manifest! Yeah! How does waiting until there is actual permanent and irreparable brain neurodegeneration make any sense once the effectiveness is proven?

From this perspective what is actually the best outcome is for the narrowest window of tau with the highest clinical benefit should be isolated for dona. This would yield the highest reported cognitive/functional benefit and the lowest p-value. The motivation here is
to say this is the line at which clinical benefit can be achieved, treat on this line or before. The label could indicate that it is not merely at the specific time of benefit but any time before.

What is probably actually scary for many on forum is realizing that many so called prodromal/Early AD were considered too advanced to benefit from dona because of their tau levels. It is remarkable to think that possibly 70% of AD patients even those with just the start of clinical symptoms might already be too advanced for help. That is something of a shocker. It is almost a non-starter to expect that early patients will say "No, I think I'll wait another year or two and then I'll be in the tau range as determined in the trial. Why would anyone do that? If they were off on their timing, they might pass through the treatment window! That makes no sense. Why wouldn't they simply treat early?
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 »

CDR-sb.GIF
SUVR.GIF
ADAS-COG12.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 »

Demo.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 »

Very exciting!

Another data point?
This is another nominally unsuccessful anti-amyloid.
Yet, the real point of interest is not the p-value versus placebo but the fit on the regression line.
That is what seems most relevant.
Once the regression line is firmly established, all that needs to happen is to move to the left and capture more of the
cognitive benefit.

[Edit: Does not appear that the SUVR for the high dose Milds is actually given.]

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6146627/
NF52
Support Team
Support Team
Posts: 2772
Joined: Tue Oct 25, 2016 9:41 am
Location: Eastern U.S.

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

Post by NF52 »

J11 wrote:Now that I have data massaged the aducan, ganta and Ban results what are some comments that can provide some context? What do all of those regressions mean? ...

From what I can tell, [aducanumab, gantenerumab and BAN2401] all follow the same regression. My take on this is that perhaps removing amyloid anywhere except at the monomer stage will cause the other amyloid species to equilibrate....

It really is reassuring to have such good agreement with the three anti-amyloids, though for things to move forward new and even more powerful clinical research techniques (such as those used with dona) need to be allow to evolve... It is important to create the space needed for innovation to occur.
... treating early should be just as helpful. The only difference is that the benefit to treatment will be implicit not explicit.Implicit means that patients will not have to experience the neurodegeneration at all; they will have been treated before it has even become manifest! Yeah! How does waiting until there is actual permanent and irreparable brain neurodegeneration make any sense once the effectiveness is proven?
J11, Thank you for your informed analysis of the current state of Phase 2 and Phase 3 clinical trials of anti-amyloid in early treatment (after MCI or AD diagnosis) and in prevention. I hope you don't mind that I selected what I see as some excellent points to quote and highlight.

In 2019, Dr. Marwan Sabbagh, FDA Commissioner, was lead author on a summary of a 2018 FDA guidance document [with explanation of some terms in brackets]:
in the last 15 years...new classes of drugs have been developed and tested...The one constant for all of these clinical trials programs is the use of the ADAS-cog [Alzheimer's disease Cognitive Scale-cognitive subscale] as the primary scale to determine efficacy... Important in their draft guidelines is the articulation of a three-stage system for classifying early AD, reflecting the pathobiology of AD... In stage 1, biomarkers [amyloid beta and/or tau and/or neurodegeneration on MR] results are abnormal, but people have no cognitive complaints or detectable clinical decline, even on sensitive tests (preclinical). In stage 2, subtle cognitive effects, but no functional deficits, appear (preclinical). In stage 3, people begin to have problems with some daily tasks measurable with instruments sensitive to AD stage 3 (prodromal), which corresponds with mild cognitive impairment [MCI] due to AD.
FDA position statement “Early Alzheimer's disease: Developing drugs for treatment, Guidance for Industry”

The preclinical phase is where current and upcoming trials are focused. That includes the AHEAD studies of BAN2401 in people with normal cognition ages 55+ who have normal cognition but intermediate and elevated amyloid beta, It also includes the projected study of ALZ-801 in the next year or two for people with amyloid beta and ApoE 4/4.

In an 2018 analysis, Alz Forum reviewed the 2018 Barcelona CTAD conference on these new FDA guidelines: Which Are the Right Tests to Satisfy New FDA Guidance?
These are people whose cognitive decline can be measured, but who are still early enough in disease that interventions could stem the worst of AD. In the draft guidance, the FDA says it will accept cognitive measures alone as the basis for approval in stage 2. It also says a drug’s benefit must be robust and consistent across multiple domains, and that the application will be stronger if supported by biomarker evidence. Petersen suggested mining longitudinal data to identify other early impairments, such as mild neurobehavioral symptoms, that could strengthen the case for a clinical benefit. “Perhaps we could marry subtle behavioral features with tau pathology,” Petersen said.

Of note, assessment choice and analysis are important for non-drug trial also.
In Barcelona, [Suzanne Hendrix of Pentara Corp., Salt Lake City] showed an example of how this method could help a [non-drug]clinical trial. Danone Nutricia Research recently reported that the LipiDiDiet trial of its nutriceutical drink Souvenaid missed its primary endpoint but met two secondary ones, hinting at some activity...[In contrast] Souvenaid showed statistical significance on the ADCOMS... Had the LipiDiDiet trial used ADCOMS, or another measure optimized for disease progression in MCI, it would have posted a positive result, Hendrix concluded...In studies where different outcome measures conflict, a three-dimensional analysis can show which measure more closely reflects disease progression and deserves more weight,
So keep your statistical skills sharp, J11!
4/4 and still an optimist!
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 »

NF52, thank you very much for replying.
{I will need to return for a more complete later, as I have an exciting post below that I want to make it to thread right away, though I will
comment further on your interesting remarks.}

My posts to date might not have been right up with the research, though I am getting better!
I am impressed to think of the level of expert insight that is now possible with internet access and determination.
I remember when they used to close the libraries at 2 AM; now the internet library is open 24/7! Hurray!
If you want to know but don't --> keep on reading!
Last edited by J11 on Sun Mar 21, 2021 6:08 pm, edited 1 time 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 »

Ganta dot.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 »

Another interesting one.
This is for a Ganta clinical trial.
What I find of note here is the really low internal correlation (0.22).
When we used the overall trial results expressed as single dots for each trial everything lined up extremely well. Yet, now when we look inside the average result we see a large amount of variation. In fact, at the trial level it seemed that perhaps the amyloid clearance and cognitive relationship topped out at ~.30 SUVR, though in this individual level regression the relationship appears to extend out to ~0.70!? This is individual level presentation (which shows the somewhat discouraging sea of almost random dots view) was what emphasized in the Statistical report of the Briefing Document.

The dona phase 2 offers us a way out of this confusion. What happens when we have an axis for amyloid and an axis for tau and an axis for CDR-sb? dona suggests that everything can move into focus. With this analysis we approach the potential of personal medicine.
A reasonable estimate might be given not based upon the group average (cloud of dots) but on an accurate individual level regression based upon amyloid, tau and possibly other markers. A much clearer separation of responders and non-responders would be plausible.

The patients in the top right of the above figure are especially prominent potential outliers. That second dimension of sortation could bring things greatly into focus. The dona results suggest that 75% of the patients could be removed due to tau levels leaving patients mostly similar to those below ~2 in the figure above. Such selection would show substantial treatment effect.

The equation in my figure above is very similar to that reported in the original article figure. I calculated the slope for one of the regressions that I posted above and for our 78 week line (versus 104 weeks for this new figure) as 0.007.
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 »

302 CDR-sb.GIF
ADCOMS 1.GIF
ADCOMS 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 »

I am somewhat grumpy that I missed the call on CDR-sb conversion to ADCOMS on 302, from the FDA Briefing Book it should have been obvious. The Memory and personal care sub-domains were both underweighted and they only ones that were not in the 23-25% decline range (Memory 28% high , personal care 15% low); so it all balanced out.

The CDR-sb takes a simple sum of the various sub-domains as shown in the first figure above. Add up 4 7s, an 8 and a 3 and you have 39. ADCOMS does better by making it more like a g factor score - i.e., add up the weighted value based upon the factor loading onto g of each of the components. We can do this from what we see in the FDA Briefing Document for the CDR-sb and we have 0.402. This is what CDR-sb would have been if it used the ADCOMS weightings (slightly higher than the reported top-line from 302 high dose). So, probably the MMSE items for time orientation or drawing brought it down slightly. As a guess we should have expected the ADCOMS estimate to be fairly close to the CDR-sb.

The 301 high dose ADCOMS apparently was a fair ways different from the reported CDR-sb. Reporting in the FDA Briefing Document would have offered some insight into the discrepancy. It would be for the best to include as much as the number crunching as possible in these reports.


Consider the strongly weighted items from the tests. With MMSE, orientation to time but not place is strongly weighted. The idea is that someone could live in the same town their whole life and even with considerable amount of cognitive impairment would probably know what town they were in. With time, one needs to be somewhat more engaged with reality to get it right. If you're not paying attention you might start to lose track of the days, weeks, months, years?? Time Orientation is strongly weighted.

Delayed word recall and orientation are also strongly weighted on ADAS-cog. Keeping track of memory type items relates strongest to ADCOMS in ADAS-COG. With CDR-sb what is particularly interesting is that now they add in hobbies and community affairs and these are strongly weighted items. What is likely speaking to is the social intelligence hypothesis (i.e., it takes a considerable amount of cognitive engagement to effectively cope with other people and complex new activities such as hobbies).

I am unclear whether AD psychometrics truly has reached the state of the art even now. I included the CDR-sb as the base measure of AD progression largely because it was often used in recent studies. However, the ADAS-cog and ADCS-ADL have also been reported in many of these studies and sometimes they have been able to report a significant result even while the CDR-sb was null.
The fact that in the aducan 302 a 40% reduction in ADCS-ADL p=0.0006 was reported versus 22% in CDR-sb p=0.0120 suggests that the ADCS-ADL simply might be a better measure. If as seems likely given all the previous posts that aducan is effective, then wouldn't
the test that gave the most emphatic YES as measured by the p-value be thought of as the more reliable? It was afteral able to give the right answer. With ADAS-cog and ADCS-ADL, they seem to have been able to do this somewhat more consistently than the CDR-sb. An even further extended ADCOMS might be able to capture yet more of AD cognitive evolution. Throw in all the aducan measures: NPI, caregiver burden, ADAS-COG ... see what happens. They are all adding to the mix.

Caregiving burden is a highly relevant aspect of the dementia experience as seen from the perspective of the dementia patient. Why not try and agglomerate all these dimensions into a composite score? Of course, if this had been done with aducan the primary result would have been even more positive.

Yet, the problem is that the top line primary result winds up taking all the weight even when there are a range of other complexifying measures that can then often be ignored. It is not that ADAS-COg or ADCS-ADl are wrong, but that for many it is Too Much Information. If people don't want to reach out for this complexity just sum it up into a more comprehensive top line.

As we have seen neuropsych symptoms such as hallucinations can be somewhat independent of cognitive scores and yet the neuropsych symptoms can be particularly difficult to cope with. Perhaps neuroimaging will be able to create a better progression predictive model that is able to include such progression features into a metric. The dona results spoke of how different brain regions responded to treatment. Neuroimaging could then be trained by the dataset to be an objective multidimensional measurer of various aspects of AD progression.
Post Reply