Somehow my last post got compressed. My middle point was that patients with a Time since Diagnosis of 6 mo (N=108) have a Hazard ratio of 0.37 (p=0.01) and patients with a Smoking Status of Never/past (N=130) have a Hazard ratio of 0.48 (p=0.02).
As far as I know this is new information and it's huge!
I did notice that the one group who fared poorly, and quite poorly at that, was current smokers as the hazard ratio was 1.61. The smoking group comprised a total of 42 patients, 17 of which were treated with Ganetespib. The phase III design will no doubt exclude current smokers. It would be interesting to see what the hazard ration and P values would become if smokers were excluded in the current data set.