For a graph of these numbers, see slide 20 of the ESMO presentation. A picture is worth a thousand words.
In the case of the 6 months since diagnosis group, the D+G has had only 7 deaths compared to 18 deaths in the D group, resulting in a Kaplan-Meier overall survival @ 11 months of .71 for D+G compared to .26 for D (p=0.010).
Synta has the option of restricting Phase 3 to this population. My guess is it will.
I don't think anybody was expecting these hugely advantageous exclusion opportumities (including the one in Slide 31 which I forgot to mention), only one of which results in a strongly statistically significant p of 0.010. The p will only get better (lower) with a larger Phase 3 N of around 500 and perhaps more restrictions.
This is extremely good news that I doubt anybody was expecting.