Especially since, according to the abstract: Of the 121 patients randomized, no significant differences were seen in baseline characteristics (gender, age, ethnicity, ECOG or disease stage.) So although the control arm enrolled very sick patients, so did the treatment arms.