Time to get back to AD! Great thing is that when you really know what AD is like, everything else is pretty much a zero. It is great being able to frame life in this way.
This could be a fairly long wait for the leca adcomm. There is not that much out there to talk about related to the submission; I thought perhaps things might need to go silent for a while -- but no! There is the phase 2 leca clinical trial that was published not so long ago.https://pubmed.ncbi.nlm.nih.gov/33865446/
With all of this time we can read through this very carefully. Typically, there is a tendency to sort of skim through readings and then often miss out on some quite important points of interest. For example, here even in the abstract there is the clarification that the primary miss for leca (the 64% instead of 80% probability of being better than placebo by 25%) was for the 12 month
The 12 month read-out? What 12 month readout? Isn't the readout meant for 18 months? Apparently not! Goal posts have been moved! Success is now defined as an 80% probability of being 25% better than placebo at the 12 month mark? 64% missed the success. Hmm, what was the probability of being 25% better than placebo at the 18 month mark? Well I had to read through for that one. It was only 76%! Another miss ... by 4%. To be clear this was Ed90 dose (i.e., the highest dose of 10mg/kg biweekly). Oh, I almost forgot the probability of straight superiority to placebo at 12 months was 97.6% (n=143)-- and as we have seen the leca phase 2 has a very high correlation to a linear regression line.
It should not be overlooked that due to the regulatory intervention to prevent APOE's from continuing to enrol in the highest dose that there was a weighting to the APOE non-4s in this highest dose: split was 107 non-e4, to 45 e4 as seen in Supplementary Table S16 which for whatever reason is not equal to the Table 2 finding of 143. Of further note is that the non e4s on the highest dose did not demonstrate much benefit. They showed a 0.011 benefit on benefit on ADCOMS (n=107) versus the baseline placebo of 0.146 (n= 70) with a 63% probability of numerical superiority to placebo and a 29% of clinical superiority (i.e., 25% difference) to placebo.
It is the APOE e4s that really carry the load in the dose result. They show a massive 0.084 benefit on ADCOMS (n=45) versus the placebo (n=168) baseline of 0.180 with a 99.2% probability of numerical superiority to placebo and 93.6 % probability of clinically significant benefit versus placebo. This isn't going to even be close in the phase 3. The highest dose used in the phase 2 was the dose chosen for the phase trial. Yet, even the second highest dose of 10 mg/kg monthly in the phase 2 in particular the APOE e4s also showed strong results for the probabilities (95.5% numerical,67.9% clinically significant with (n=218) for treatment and (n=168) for placebo).
The supplementary appendix runs through various similar results for ADCOMS / CDR-sb / ADAS-cog 14 for Bayesian and MMRM at 12 and 18 months. The ADAS-cog 14 Bayesian result in Table S8 appears to be especially strong. I am not entirely clear why the results do not appear much improved at the 18 month versus the 12 month mark.
There should be ~ 1200 patients who have been in the phase 3 now for at least 12 months. Table S4 notes that the 10 mg/kg biweekly dose with n=152 showed a p-value of 2.7% on ADCOMS versus the placebo (n=238) at 12 months. It would then be highly instructive if the rolling submission were to include an interim glimpse of how the 1200 patients might be doing. Prespecifying APOE e4 might be all the more instructive (remember that only 45 e4s were needed to establish 99.2% probability of numerical superiority). If such analysis were to be included in the FDA Briefing Document, then that would remove most of my enthusiasm from posting to the thread. At some time in the not too distant future questioning the efficacy of anti-amyloids (especially lecanemab) might no longer be felt productive. This is not even going to be close.