AVEO didn't breakdown number from that 30% patient population, based on other numbers provided, my guess is the number didn't differ from Nexavar that much.
The wording in the PR (#msg-70454501) makes it certain that the PFS difference in the 30% subgroup in the second-line setting was not statsig, and we also know that the median PFS in the control arm was 9.1 months for the 30% subset. For the sake of discussion, say that the reported PFS medians for the Tivo arm in the full trial and in the 70% subset had instead been the reported means; then the PFS mean for the Tivo arm in the 30% subset would have been 10.2 months. In all likelihood, the median PFS of the Tivo arm in the 30% subset was not markedly different from this figure.
Off hand I seem to recall that of the $1 billion in Nexavar sales, around 80%-90% are for HCC. Jq1234 is correct that Nexavar sales growth has been almost exclusively in HCC.