Background: Transarterial chemoembolisation (TACE) is the standard of care for patients with intermediate stage hepatocellular carcinoma, while the multikinase inhibitor sorafenib improves survival in patients with advanced disease. We aimed to determine whether TACE with sorafenib improves progression-free survival versus TACE with placebo.

Methods: We did a multicentre, randomised, placebo-controlled, phase 3 trial (TACE 2) in 20 hospitals in the UK for patients with unresectable, liver-confined hepatocellular carcinoma. Patients were eligible if they were at least aged 18 years, had Eastern Cooperative Oncology Group performance status of 1 or less, and had Child-Pugh A liver disease. Patients were randomised 1:1 by computerised minimisation algorithm to continuous oral sorafenib (400 mg twice-daily) or matching placebo combined with TACE using drug-eluting beads (DEB-TACE), which was given via the hepatic artery 2-5 weeks after randomisation and according to radiological response and patient tolerance thereafter. Patients were stratified according to randomising centre and serum α-fetoprotein concentration (<400 ng/mL and ≥400 ng/mL). Only the trial coordinator was unmasked to treatment allocation before patient progression during the study. The primary endpoint was progression-free survival defined as the interval between randomisation and progression according to Response Evaluation Criteria In Solid Tumors version 1.1 (RECIST v1.1) or death due to any cause, and was analysed by intention-to-treat. Safety was analysed by intention-to-treat. The trial has been completed and the final results are reported. The trial is registered at EudraCT, number 2008-005073-36, and ISRCTN, number ISRCTN93375053.

Findings: Between Nov 4, 2010, and Dec 7, 2015, the trial enrolled 399 patients and was terminated after a planned interim futility analysis. 86 patients failed screening and 313 remaining patients were randomly assigned: 157 to sorafenib and 156 to placebo. The median daily dose was 660 mg (IQR 389·2-800·0) sorafenib versus 800 mg (758·2-800·0) placebo, and median duration of therapy was 120·0 days (IQR 43·0-266·0) for sorafenib versus 162·0 days (70·0-323·5) for placebo. There was no evidence of difference in progression-free survival between the sorafenib group and the placebo group (hazard ratio [HR] 0·99 [95% CI 0·77-1·27], p=0·94); median progression-free survival was 238·0 days (95% CI 221·0-281·0) in the sorafenib group and 235·0 days (209·0-322·0) in the placebo group. The most common grade 3-4 adverse events were fatigue (29 [18%] of 157 patients in the sorafenib group vs 21 [13%] of 156 patients in the placebo group), abdominal pain (20 [13%] vs 12 [8%]), diarrhoea (16 [10%] vs four [3%]), gastrointestinal disorders (18 [11%] vs 12 [8%]), and hand-foot skin reaction (12 [8%] and none). At least one serious adverse event was reported in 65 (41%) of 157 patients in the sorafenib group and 50 (32%) of 156 in the placebo group, and 181 serious adverse events were reported in total, 95 (52%) in the sorafenib group and 86 (48%) in the placebo group. Three deaths occurred in each group that were attributed to DEB-TACE. Four deaths were attributed to study drug; three in the sorafenib group and one in the placebo group.

Interpretation: The addition of sorafenib to DEB-TACE does not improve progression-free survival in European patients with hepatocellular carcinoma. Alternative systemic therapies need to be assessed in combination with TACE to improve patient outcomes.

Funding: Bayer PLC and BTG PLC.