ATL - Here is my prior post. It was before we learned of the new endpoints, so the numbers reflect combining new and recurrent GBM:
FWIW - Some back of the envelope calculations to estimate unblinded 3 year OS in the DCVAX ITT group using very conservative assumptions. We don't have info to estimate 5 year OS.
These calculations are based on the following:
1) Overall blinded median OS from the JTM article is 24.2 months(based on those actually in the trial for 36 months at the time of JTM publication)
2) Assume all crossover patients did exactly as well as those assigned initially to DCVAX
3) The 30 control patients who never crossed over live to the standard of care median OS of 17 months.
Using this input, the median 3 year OS for the ITT DCVAX group alone increases to 25 months (from 24.2 months in the combined group). This tells us that if DCVAX works just as well for rGBM as nGBM, what the JTM article reported likely represent a fairly accurate picture of what median 3 year OS would be in the unblinded DCVAX ITT group (i.e., splitting out the controls does not really change things much). This is the lower limit of 3 year OS for the ITT DCVAX group that we could expect in the TLD.
What is important to recognize though is that these unblinded DCVAX ITT median OS numbers increase further to the extent that crossovers (rGBM) do not do as well as those originally treated with DCVAX (which to me does not seem to be an unreasonable assumption). Just as an example, if we assume that unblinded 3 year median OS in the crossover group was 20 months (rather than 17 month SOC), the unblinded median OS for the ITT DCVAX group becomes 26.8 months (an increase of 2.6 months over the blinded JTM data). This to me seems to be a pretty realistic assumption. Regardless, it doesn't seem that unblinded median 3 year OS will increase dramatically from what has been reported based on blinded data unless DCVAX simply fails to work in rGBM. Caveat - Everything above does not address what 5 year median OS will look like on an unblinded basis, or what the KM analysis would show.
