The purpose of this study is to determine whether RADIOEMBOLIZATION with 90 Yttrium microspheres is more effective on overall survival in advanced Hepatocellular carcinoma (HCC) with or without portal venous obstruction and no extrahepatic extension than sorafenib which is now the standard treatment of advanced HCC.

Background: In patients with advanced hepatocellular carcinoma, sorafenib is now the standard treatment with an increased median overall survival but an overall incidence of treatment-related adverse events of 80%. There is growing interest for RADIOEMBOLIZAION with 90 Yttrium microspheres. It involves infusion of embolic microparticles of glass or resin impregnated with the isotope yttrium-90 through a catheter directly into the hepatic arteries. A substantial number of open-label single-group studies showed supporting evidence for a potential efficacy on overall survival and acceptable or low toxicity. Trial design: multicenter, prospective, controlled, open label randomized trial of Y90 RADIOEMBOLIZATION versus sorafenib. Participants: Adult patients with 1) advanced HCC according to BCLC staging system (stage C) with or without portal vein thrombosis 2) ECOG performance status of 2 or less 3) adequate haematological, renal and hepatic functions 4) liver cirrhosis Child Pugh A - B7 and 5) no extrahepatic metastasis. Interventions: In the sorafenib group, patients will receive continuous oral treatment with 400 mg of sorafenib twice daily. In the Y90 RADIOEMBOLIZATION group, patients will first undergo angiography and scintigraphy for eligibility assessment (absence of or acceptable lung shunting) and preconditioning (embolization). RADIOEMBOLIZATION therapy with infusion of Y90 microspheres will be performed secondly. Objectives: The primary objective is to compare the efficacy of Y90 RADIOEMBOLIZATION to sorafenib in the treatment of advanced hepatocellular carcinoma. Secondary objectives include the comparison of safety profiles, quality of life and health care costs between the two therapeutic groups. Outcomes: The primary endpoint is the median overall survival time. Secondary endpoints include adverse events reported according to the NCI CTC, progression-free survival at 6 months, response rates, general and hepatic-specific quality of life scores, health care costs which comprise the MICROCOSTING of Y90 RADIOEMBOLIZATION from the viewpoint of the hospital and the full cost of each strategy. Sample size: 400 participants (200 par arm). The trial have 80% power to detect a clinically meaningful increase in median survival time of 4 months between sorafenib (expected median survival time 10.7 months) and Y90 RADIOEMBOLIZATION (expected median survival time 15 months) with a two-tailed type I error risk of 5%. Randomization: 1 to 1 randomization will be stratified according to recruiting center, ECOG performance status (a score of 0 vs. a score of 1 or 2), presence or absence of macroscopic vascular invasion (obstruction of portal vein or any branch vs none) and previous chemoembolisation failure . Randomly permuted blocks of random sizes will be used. Study duration and Setting: Accrual period 24 months. Additional follow-up period: 12 months. 14 centres involving both clinicians (hepatologists, hepatobiliary surgeons, and oncologists) and radiologists and Nuclear medicine physicians on each site.

Inclusion Criteria:

• Histological or cytological diagnosis or meet the AASLD criteria for diagnosis of HCC and at least one uni-dimensional lesion measurable according to RECIST criteria by CT-scan or MRI
• Adult over 18 years old and estimated life expectancy over 3 months
• Patient with advanced HCC according to BCLC staging system (stage C) with or without portal vein thrombosis, not eligible for surgical resection, liver transplantation nor radiofrequency ablation OR patient with progression or recurrence of HCC after surgical or locoregional treatment not eligible for surgical resection, liver transplantation nor radiofrequency ablation OR patients in whom chemoembolisation has failed after two courses (patients who have received only one course of chemoembolisation are eligible if the failure of the first round shows that a second round will have no more impact; patients who have received more than two courses of chemoembolisation are still eligible if the arterial network is perfectly normal on a CT scan in the arterial phase). Failure is defined as the absence of an objective response after two courses of treatment in the treated nodule (objective response according to modified RECIST criteria and/or EASL criteria).
• ECOG performance status under or equals 1
• Adequate haematological function: Hb over or equals 9g/100mL, absolute neutrophil count over or equals 1 500/mm3, platelet count over or equals 50 000/mm3
• Adequate renal function; serum creatinine under 150μmol/L
• Bilirubin under or equals 50 µmol/L, AST or ALT uner or equals 5 x ULN, INR under or equals 1.5
• Liver cirrhosis Child Pugh A - B7
• written informed consent

Exclusion Criteria:

• Another primary tumour, with the exception of conventional basal cell carcinoma or superficial bladder neoplasia
• Extrahepatic metastasis
• Advanced HCC previously treated
• Advanced liver disease with Child-Pugh score over 7 or active gastrointestinal bleeding or encephalopathy or ascites refractory to diuretic therapy Women who are pregnant or breast feeding
• Allergy to contrast media
• Contraindication to hepatic artery catheterisation, such as severe peripheral vascular disease precluding catheterisation
• Psychiatric or other disorder likely to impact on informed consent
• Patient unable and/or unwilling to comply with treatment and study instructions
• Patient unable to swallow oral medications