Sentimentstocks did some good DD on the vanguard group. Blood sweat and tears trying to read all those lines and steps.
38 in that vanguard (2008/2009) group if I recall.
38 total unidentified mgmt as well in the entire trial population of 331.
Coincidence? Maybe so.
24 unidentified mgmt amongst 35 noncrossovers
14 unidentified mgmt amongst 296 treated.
Your theory: High number of unidentified that didn’t make it long enough to crossover (assumption from DD, vanguard group didn’t due well) (German Colnel thinks noncrossovers did well) unresolved.
One way to organize the vanguard would have been to censor, another would simply be to exclude them in the new endpoints. Perhaps both were done.
Still, 14 unidentified mgmt in current endpoint groupings.
3 in rGBM out of 64 (4.7%)
11 in nGBM out of 296 (4.7%)
That’s not enough to really disrupt the trends upwards or downwards. Some of us back in the day would have guessed downward but today German Colnel would argue upwards.
Hmmm. The noncrossovers would definitely be impacted at 63% unidentified mgmt. These are the ones that you mostly assume couldn’t survive long enough in the early days to make it to crossover treatment.
Hmmm. As You probably read earlier, I did think unidentified mgmt were likely to have been mostly from the earliest cohort, so I agree with you there.
My memory is cloudy, however regarding how long they made patients wait in the vanguard cohort to cross over. Hmm.
Logic dictates
1. you’d want to crossover,
2. Most mgmt unidentified did not crossover
3. Most mgmt unidentified likely came from earliest cohort.
4. Senti’s and our DD from years ago thought vanguard group survival not as strong.
The logic starts to get circular, but each leg is part of the chair.
Hmmm.
AI
Market Data 
Markets 
