With regard to ARQL, the question I'm grappling with is why was there such a big difference in OS between the placebo arm in the brivanib BRISK trial and the ARQ197 2nd-line HCC trial? Both trials enrolled pts who had progressed on, or were intolerant to, sorafenib but had good liver function (Child-Pugh A) yet the OS in the tivantinib control arm was only 3.8 months as compared to 8.2 months in the brivanib trial. Caveats about direct comparisons between trials aside, the 7.2 month OS tivantinib achieved in the 2nd-line, high MET population isn't nearly as compelling when one considers that even with an OS of 9.4 months brivatinib ultimately failed to meet its primary endpoint. Any help understanding this issue would be greatly appreciated.