Thanks for your delayed comment on my potency assay suggestion regarding the change in IL2Ralpha inhibition.
I do believe this particular dataset, which was available at ODAC, only contained the GVHD001 study results in line with the primary end point ( Only went out to 28 days ). I'm not really quite sure why the company chose 28 days to highlight this correlation, other than to keep in timing with the primary end point.
As expected, if you expand the data set out over time ( more time ), you can see there is a much greater effect over time as the patients continue to improve.
The top graphs are graphing similar but not identical combinations of T-Cell activations, as the 28 day IL2Raplha vs t-cell activation graphs JB repeated.
JB - where exactly is the missing information they have " patched " in the in vivo vs in vitro graph ? My take is they only used the information that is in the graph to generate the scientificly meaningful relationship, P = 0.04. Or are you saying because the data was not available for all 51 patients, it therefore is patched?
In any case, the end result speaks for itself.
Day 180:
Overall ( All measured patients ) 85% survival vs 54% P = 0.01
MAP > 0.29 = 100% survival vs vs 17% P = 0.003
Minnesota high risk = 89% survival vs 50% P = 0.01
Grade D disease = 91% survival vs 50% P = 0.03
Firstly - as the FDA have themselves indicated the MAP score appears to be the most accurate measure of the likelyhood of mortality there currently is. That is evident in the 17% survival for patients recieving cells that had lower inhibiting activity, than say grade D or Minnesota high risk who were 50% recieving the same cells with lower inhibiting levels. If you look at the High MAP group, who recieved the better cells, they have a an even higher survival rate than the overall population - that is the opposite of what the map is indicative of. As you can see high MAP + high inhibiting cells = 100% survival, vs High + Low MAP + good cells = 85% survival in the general population.
This contradicts the very meaning of a high MAP score patient, however is explained by the MOA of the cells being activated in the presence of inflammation. The more the inflammation ( MAP score ), the more activated the cells are, and we can see by the data, the mortality rate is reduced below that of the non high map group.
If IL2Raplha inhibition levels in vitro were actually linked to survival, then you should see a greater survival level in patients who had higher in vitro inhibiting levels. Dr Kurtzberg has already demonstrated this above. That is exactly what occurs in every single group. In the whole group, in the High Map group, in Class D group, and in Minnesota high risk group the exact same relationship is present, and at scientifically relevent P values. For this to be achieved in such a small sample size means the effect level must be high.
The FDA themselves were the ones that made a huge song and dance about the potency assay ( TNFR1
levels ) not correlating to survival or day 28 overall response, or anything in vivo. The company is now able to show that IL2 levels in vitro from the other assay, clearly do scientifically correlate with
much higher survival in all groups.
How did the GVHD001 trial result in primary end point achieved, when the other trials did not? They changed the acceptance criterea for the TNFr1 levels.
You are asking, if it's valid to " tighten" the acceptance criterea again, this time for the in vitro IL2raplha inhibition assay. The statistics are all scientifically relevant - so why do you think statistically this is not valid?
No data has been discarded, and no data has been added.
The company also has correlation from 280 and EAP 275 from single lot donor patients, that also correlates the same thing, and that is lower IL2 inhibition is linked with a lower overall survival. All the data, all the statistics are saying the same thing here? You can't just throw out a " I don't think this is statistically valid" comment and walk out the door. Statistically it has met the definintion of being scientifically relevant.
(20min delay)