Broiler_x, I will take your comments line by line.
I think this is another poor decision by the FDA, but not as bad as for Aduhelm.
I do not think the FDA realistically had any other option then some variant of approval. It might have put a black box warning for tPA, or restriction on anti-coagulant use, though on the topline yea or nay for lecanemab I really do not see a way to nay. This will likely be verified as other health authorities decide on this. It is not easy to see them rejecting it. When the FDA did not even consider that an ADCOMM was called for that probably should have given us a clear indication of where this one was going.
I also don't expect payers (insurers and Medicare, etc.) to pay for this drug due to it's very marginal efficacy data.
Yes, this does seem to have been the correct call on coverage. CMS somewhat surprisingly apparently wants to maintain its non-coverage CED policy from Aducanumab through to full approval. This will be very difficult for millions of people with Alzheimer’s dementia. A submission for full approval was made the same day as Lecanemab was granted accelerated approval. This could become quite exciting because they might call for an ADCOMM and then we can see all the details about side effects, etc. and have much deeper insight into various aspects of Lecanemab. Without the FDA briefing document the approval process was much much less transparent to me. As an initial guess, it wouldn't be too much information -- it will be more that more information will reveal responders etc.. This will be great!
I have been studying this antibody very extensively over the last few months and I've discussed these results with several AD experts. We all agree: like Adulhelm, the biomarker data looks impressive and unlike Adulhelm, there does appear to be a reduction in cognitive decline at the beginning of dosing.
My take on the cognitive benefit is that Lecanemab is actually quite similar to Aducanumab. On the Aducan thread I have done nearly non-stop regressing and my conclusion is that Aducanumab, Lecanemab and Gantenerumab are all on the same regression parabola. This is somewhat surprising.
Here is the parabolic regression. Aducan was part of the training set. It is truly remarkable how closely Clarity, Graduate 1 & 2 conformed to this prespecified equation. By simply inputting the amount of amyloid removed on PET SUVR one can nearly perfectly determine the cognitive benefit at the group level. This is the basis that the FDA was using for accelerated approval. When the numbers are out by only ~0.02 CDR-sb, it becomes almost impossible to deny that amyloid removal is a valid biomarker. Strangely, amyloid removal as a biomarker then becomes a much stricter restriction than merely hitting a p-value. With the regression it is not good enough to hit any ole p-value; it becomes important to hit a specific p-value that corresponds to the regression line. I have no idea how one could go about setting up the numbers for that.
https://www.desmos.com/calculator/v3uag5xlcm
The PET amyloid SUVR is now thought not to be the actual disease target of interest, though as long as the mabs reduce the down stream beta amyloid as plaques, then an identical cognitive benefit arises. So, the only reason Lecanemab has superior results to Aducanumab and Gantenerumab is because it removes more amyloid. Moving out even marginally to the left on the x-axis to more amyloid clearance gives quadratically improving cognitive benefit.
However, that benefit appears to go away after 6-9 months. After that, with continued repeated dosing there is no difference between the placebo and the drug in reducing cognitive decline.
When I datagrabbed the results from Clarity, I found that the difference between Lecanemab and placebo continued to grow (linearly) right up to the 18th month and likely beyond.
Drag the b slider equation 25 for the vertical and H slider equation 41 for the horizontal slider to see the differences in Clarity. With CDR-sb, the treatment and placebo arms were very very linear. The R^2 in equations 7, 12 and 17 are all nearly 100%. You have a good feeling extrapolating when it is that solid. By eye it is not easy to detect a treatment effect. Yet, when you do a linear regression, the benefit clearly grows through time.
https://www.desmos.com/calculator/8ezgghsdpm
ADAS-cog was less linear, though when you use the centroid of the data using the linear approach you wind up with a 36% benefit which is right in line with other trials have found. Without the linear regression you just have all these jumpy points that seem to have a lot of bounce to them. The headline number that was then reported for Clarity of 26% reduction in decline does not seem as believable or accurate. ADAS-cog is an objective cognitive test so this likely has more forward predictive value than would the CDR-sb which is more retrospective in orientation.
https://www.desmos.com/calculator/haccvoordl The CDR-sb's 27% topline then seems less in tune with the actual benefit in the trial.
This was seen with Aducanumab. Even after years of dosing those on treatment appeared to continue to have a better result relative to placebo. There could be selection involved, though from current evidence it does appear that the benefit continues to increase.
Keep in mind, this also does not stop or reverse cognitive decline, but just reduces it for a brief period.
From the recent phase 2 articles that you referenced I determined the percentile distributions of the different dosing arms. Can't link to the figure directly though it is the Fri Jan 06, 2023 2:02 pm post on the Aducan Celebration thread. The claim that has often been made that treatment does not stop or reverse cognitive decline is not congruent with this evidence. Even 20% of the placebo were stable or improved! For the high dose treatment arm in the phase 2 it was ~40%. Of course, 20% of the placebo improving is not believable; true AD patients do not improve over 18 months, by definition they must decline. However, this is more a problem of how the patients were selected in Clarity. In Trailblazer 2, they used dual tau and amyloid selection and this should amplify the treatment effect signal as it did in Trailblazer (the original). In the most selected group this yielded a 50% reduction in decline on iADRS.
This will be the next news to break when the TB2 completes in April. Moving the ball upfield from 27% reduction in decline to 50% might ruffle some feathers.
The paper with the clinical results is available freely here:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9768996/
The data I'm referring to is Fig. 9 which shows the cognitive efficacy graphs.
I would paste that data here, but I'm not sure what the rules are about inserting data from anther publication.
The problem here is that we are comparing treated patients with treated patients. What we have seen with Aducanumab is that when we compare the patients who became amyloid negative to those that were still amyloid positive after being treated with Aducan at some point these amyloid positives experience cognitive collapse (i.e., a decline of 4 or more CDR-sb points). None of amyloid negatives exhibited this trajectory. The problem is that AD clinical trials typically only are randomized for 78 weeks. Over the longer term as AD progresses there can be larger cognitive decline that was not seen in the shorter term trials.
Now, you might be thinking: reducing decline for a few months is still great and better than nothing! I agree, however it comes at significant costs.
From my perspective on AD, I feel that Lecanemab represents the end game of dementia in my family. With us, we have experienced cognitive decline that spans decades. It hasn’t been that we did not understand what was happening. It was just there was nothing that could be given as a treatment. Lecanemab is such a relief. I know that within the next few weeks there would be some treatment that could likely be of help to me if I recognized cognitive decline in myself or other family. In our experience even 27% reduction in decline over 30 or 40 years would mean that end stage AD would probably never arise.
This one shows how the longer term benefit could accumulate. Notably here we are only talking about starting dosing at the MCI stage. With us there was at least a decade or two before the MCI stage where there was clear cognitive dysfunction. The potential to bend the curve over such an extended period of time would almost certainly mean that AD would no longer be a force in our life.
https://www.desmos.com/calculator/qx1eugcq0l
The logic involved in going off label if needed, at a micro dose, with extreme safety, minimal cost, and a good chance for positive treatment benefit seems compelling to me. My best estimate is that this would not offer us months of improved quality of life, but probably essentially a functional cure.
First, there have been 3 treatment-related deaths in these trials. All three had to do with patients who had strokes and were given the standard of care for strokes (like blood thinners).
The safety during the randomized portion of Clarity was impressive. There was one fatality due to intracranial hemorrhage during clarity. Specifically, this was a macrohemorrhage that occurred in the placebo arm. In fact there was considerable neuropathology during the core phase of Clarity in the placebo arm. There was substantial ARIA-H microhemorrhages (7.0% of these patients experienced a microhemorrhage). I do not believe is yet widely understood that there is substantial CAA, stroke like events etc. that occur even without mab treatment. The side-effects that are occurring during the trials are not against some pristine background of health. The ARIA etc. that is occurring exactly because none of this existing pathology has ever been treated. These patients are walking around and a range of stroke like risk is simply waiting to be activated.
The response from the drug companies and clinicians is to not use those drugs with anyone who is using Lecanumab. So, if you have a stroke and you're on Lecanumab, you're just out of luck?
Yes, I do agree that this puts everyone in a difficult position. For some patients, there simply might not be a treatment option if they had a stroke and they needed tPA. Clearly that would be a very difficult clinical choice. However, this might not have been true with the Lecanemab patient who had a stroke and then was treated with tPA. In this instance treating with thromboectomy might have been a treatment alternative. The MRIs showed how devastating treating with tPA could be.
I also want safer treatment. With Lecanemab this evolution towards enhanced safety continues. For example, in the Clarity OLE they are now dosing subq. It is possible that when the FDA decides on full approval they might be able to offer a label indication for subq dosing. This would be a great victory for patients if this were possible. Hopefully, now that the finish line is clearly in sight and the funding window for full approval has been opened, that efforts will be redoubled to offer subq.
The second big cost is actually that: the cost. Pricing has not been announced, but since it is an antibody it will be expensive and likely greater than $20K/year.
I am not completely sure about $20k/year. For so many of the people on the sidelines, mab treatment could be decades into the future. They will never pay the patent price. For us, if we were to need early treatment, perhaps even 1 mg/kg treatment a few times per year would be all that we would want. For many patients in such a circumstance mab therapy could largely be an out of pocket payment.
And patients must start taking this years before showing any significant cognitive decline. And we won't even know if those patients will actually ever develop AD.
In our experience all of the family developed it. It wasn’t sporadic. I have full genomes to work with and will be able to use predictive genetics to confirm which of us are at risk. Genome prices might move down to $100 this year. Biogen spent large money on a sequencing project. If genome sequencing were to go mainstream this year at a $100 price point the entire human genome could unlock and then it would be super easy to find which variant was causing our problem. I could then go online and find others with the variant and find out if they were helped by Lecanemab. This would be an amazingly powerful way of fitting the right patient with the right genotype to the right treatment. With logic as strong as that I doubt whether I would even get push back from a doctor.
I am not sure whether such an effort would arise from the corporate side, though it would clearly be an extremely powerful and persuasive argument if one could accurately predict treatment response by specific genotypes. There are many other genotypes aside from APOE epsilon 4 that are likely important for response. If this is all 0.45 CDR-sb, that is one thing. It would be entirely different, if the superresponders could be prospectively identified.
So just start dosing anyone in their 60's or earlier who might have a family history of AD, or maybe displaying earlier MCI for the rest of their lives?
The genetic approach could probably be helpful for others as well. Many of the patients who enroll in AD trials probably do not have extremely strong family histories. Finding those patients that are core AD patients will help to select those who would benefit.
Sounds awesome for Biogen. But that would financially devastate Medicare and insurance companies. So this would dramatically raise health care costs for everyone, for a drug that might slow cognitive decline for a few months? And treat people that we don't know will ever develop AD? I don't think payers will reimburse for that.
The economic argument in favor of Lecanemab treatment for those in the pre stages might become extremely compelling. Consider someone in their 40s earning $80K with children. If that person were to descend into AD over the next 10-15 years they would go from rowing the boat to being a passenger.
The accounting would likely find the cost involved as being large. In our experience the negative externalities are also substantial. Dementing illness becomes like an internal force of destruction of society. Preventative treatment would likely have significant financial paybacks even over the short term. If this all resolves down to money, then at some point those with the sharp pencils might finally realize how massively expensive it is not to treat this. The current hard dollar cost of AD is ~$350 billion and rapidly rising; there are considerable other costs involved that also would eventually translate into real world dollars. My impression is that we have already crossed over to a time that it costs more not to spend more.
I actually understand why the FDA approved this, because they simply look at safety and efficacy and not economics and effects on health care costs. That's not their job. I also like how this has caused excitement in the AD research field which has been so dismal for so long. I'm definitely seeing a lot more interest from investors in this field which we haven't seen for years. So for that, I applaud this effort. But I don't see this as actually doing much to help people I'm afraid.