Everyone went on weekly gemcitabine as soon as they could post radiation tx. Some had erlotinib mixed in later. No data on how often.
I don't think this late chemo had any effect. Such effect should be washed out by the randomization since both cohorts had the same tx - except for the 5 weeks of TNFerade early.
Either the effect is just due to lumpy data that will smooth out in the end or its due to TNFerade. If the later, it seems to me there could be a subset of LAPC with a little less disease who respond to the necrotizing effect of the drug in the same way only those with less disease respond to a Whipple. Those with more micromets may simply not see any benefit and die with the curve before 10 months.
I hate subset explanations for data. They are lousy ideas to invest in.