Overall survival determined as the time (days) from the date of randomization at start of study to the date of death, due to any cause. Outcome measure was assessed regularly, i.e. every 3 weeks for the first 24 weeks during treatment and every 4 weeks thereafter and approximately every 3 months during post-treatment.

Evaluations based on ITT population. Subjects alive at time of analysis were censored at last date of follow-up (FU) (last visit or contact or at data cut-off date). In case of incomplete date, day was missing, day 15 was used.

Overall survival determined as the time (days) from the date of randomization at start of study to the date of death, due to any cause. Outcome measure was assessed regularly, i.e. every 3 weeks for the first 24 weeks during treatment and every 4 weeks thereafter and approximately every 3 months during post-treatment.

Evaluations based on ITT population. Subjects alive at time of analysis were censored at last date of FU (last visit or contact or at data cut-off date). In case of incomplete date, missing day, day 15 was used. Placebo censored at 30June2005, approximate time of crossover of placebo subjects to sorafenib. NA - not estimable due to censored data.

PFS determined as the time (days) from the date of randomization at start of study to the actual date of disease progression (PD) (radiological or clinical) or death due to any cause, if death occurred before PD. Outcome measure was assessed approximately every 8 weeks using RECIST v1.0 criteria by independent radiologic review. Radiological PD defined as at least 20% increase in sum of longest diameter (LD) of measured lesions taking as reference smallest sum LD recorded since treatment started or appearance of new lesions.

Evaluations based on ITT population as of 28Jan2005 data cut; 769 subjects randomized at that time. PFS determined as time from randomization to actual date of disease progression (PD) (radiological or clinical) or death, if death occurred before PD. Subjects without PD or death at time of analysis were censored at last date of tumor assessment.

Best overall response was determined according to Response Evaluation Criteria in Solid Tumors (RECIST) v1.0 by independent radiologic review. Categories: complete response (CR, tumor disappears), partial response (PR, sum of lesion sizes decreased), stable disease (SD, steady state of disease), progressive disease (PD, sum of lesion sizes increased) and not evaluated.

Evaluations of best overall response based on the valid for response population, where as per protocol, subjects were to have first post-baseline tumor evaluation performed at the end of Cycle 1 (6 weeks post-randomization). Of the ITT population that met this criteria as of the 28Jan2005 data cut, 672 subjects were valid for response.

Primary Analysis for FKSI-10 patient-reported outcome (PRO) measure defined as longitudinal analysis of mean score over the first 5 treatment cycles. FKSI-10 patient responses for each question range from "0=not at all" to "4=very much" and after reverse coding the range of values for FKSI-10 total score is from 0 to 40; higher score represents better HRQOL.

Evaluations based on ITT population with a PRO assessment. Day 1, Cycle 1 served as baseline assessment.

Primary Analysis for FACT-G (using PWB score) patient-reported outcome (PRO) measure defined as longitudinal analysis of mean score over the first 5 treatment cycles. FACT-G (PWB score) patient responses for each question range from "0=not at all" to "4=very much" and after reverse coding the total FACT-G (PWB score) range of values is from 0 to 28; higher score represents better HRQOL.

Evaluations based on ITT population with a PRO assessment. Day 1, Cycle 1 served as baseline assessment.