Oren, please re-read gr8's comment. If Ph2 VB111 showed:
VB-111 Ph2 results:
54% - 57% @ 52 weeks*
50% @ 59 weeks
Avastin's historical data suggests:
50% @ 34 weeks**
28% @ 52 weeks ***
* Both 54% and 57% are used in presentation. (Slide 12 & 22, Sept17 presentation)
**I took average after removing the best & worst data points from historical control data. (Slide 21, Sept17 presentation)
*** I took the average after removing the best & worst data points from historical control data. (Slide 22, Sept17 presentation)
Now let's look at your assumption, Oren. If Avastin and Avastin+VB-111 are performing similarly at mOS, then that would imply either of these two scenarios:
Scenario 1: Avastin is MAJORLY over-performing.
This assumes no change in VB-111 performance, so Avastin alone needs a 52-week survival around 54-57%. I have to hope this high unlikely event did not happen.
Scenario 2: VB-111 in combo with Avastin is MAJORLY under-performing.
This assumes no change in Avastin performance, so Avastin+VB111 must have a mOS of closer to 34 weeks. Then, your assumption about 12-month survival being different between treatment and control would then require the surviving Avastin patients continues to diminish to 28% @ 52 weeks. However, the long-tail of survival due to VB111 will manifest as a survival rate somewhere between 35-50% @ 52 weeks for Avastin+VB111.
This is not out of the realm of possibility. I would be a tougher road to approval, because things would become a bit more subjective. Dror in recent talks has mentioned looking beyond just mOS (ie. looking for long-term survival relative to what is known with other IO agents, correlating response to initial tumor size, etc).
Oren, is Scenario 2 what you are saying is currently happening?
Given the rigor of Ph2 trials and that it did include VB111+Avastin, I would like to think that VB111+Avastin is performing closer to what was observed in Ph2. And so I would expect both the mOS and long-term survival to lean significantly in favor of VB111.
