Handicapping Satraplatin accelerated approval:
>With all Dew respect - what's your estimated odds for approval?<
I assume you are asking about accelerated approval, specifically, based on only the PFS data. I see the following cases (all figures are JMHO, FWIW):
1) FDA asks ODAC (or ODAC asks itself): “Should FDA wait a few months in order to see the overall-survival data?” I.e. the survival data are a focal point of ODAC’s evaluation.
• 1a) ODAC says: “Yes, let’s wait for the OS data.” Probability of accelerated approval is 16%.
• 1b) ODAC says: “No, don’t wait for the OS data—you have enough to approve without it.” Probability of accelerated approval is 88%.
2) The question in #1 above is not asked; i.e. the OS data are not a focal point of ODAC’s evaluation.
• 2a) ODAC finds the PFS data compelling. Probability of accelerated approval is 72%.
• 2b) ODAC finds the PFS data not compelling. Probability of accelerated approval is 4%.
--
How to handicap the cases themselves? I would say that cases 1a, 1b, 2a, and 2b have probabilities of 48%, 26%, 19%, and 7%, respectively. Then the probability of accelerated approval is:
.48(.16)+.26(.88)+.19(.72)+.07(.04) = 45%.
If we know that OS is in fact a focal point of the ODAC analysis (case 1), then the probability of accelerated approval is:
(.48/(.48+.26))(.16)+(.26/(.48+.26)(.88) = 41%.
If we know that OS is not a focal point of the ODAC analysis (case 2), then the probability of accelerated approval is:
(.19/(.19+.07))(.72)+(.07/(.19+.07)(.04) = 54%.
--
Obviously, the probability of accelerated approval would be very much better if the OS data were not almost mature. Once again, all numbers are JMHO, FWIW
Market Data 
Markets 
AI
