vmlg17,
I am pretty sure their estimates are based on assumptions. Just taking the alive patients and assigning a 2:1 treatment to SOC only ratio with 331 patients gets you pretty close. For example and argument's sake only, in the most conservative estimate treatment = placebo. Then we can say that when the 248th PFS event hit that because SOC and treatment in this example are expected to have equal effect, the remaining patient ratio must be nearly the same as the beginning ratio of treatment to SOC only patients at 2:1 and that 1/3 (SOC only arm) of the remaining 83 non PFS event patients = ~28. From here we see that these remaining 28 SOC only patients as a % of the total patient population is 28/331= ~8.5%. This would mean that the remaining 91.5% of the patient population would be equal to the total patient population expected to have received treatment at some point. When modifying for lost to followup and other potential factors the numbers get adjusted accordingly which is what they gave us. Best wishes.