Most patients with hepatocellular carcinoma present late with unresectable disease and, to date, an effective and safe palliative treatment has not been found. From December 1971, to December 1981, a randomized controlled trial was conducted to evaluate the following treatments against a control group receiving only symptomatic therapy (group I): hepatic dearterialization (group II), hepatic artery ligation and cannulation for infusion of chemotherapeutic agents (group III), hepatic artery ligation and portal vein cannulation for chemotherapeutic agents infusion (group IV), and external irradiation (group V). Response to various treatments was assessed with respect to symptomatic relief and duration of survival. A total of 166 patients with histologically confirmed tumor confined to the liver were studied. All patients had similar preoperative general status and biochemical findings. Patients in group IV had the best symptomatic improvement; 58.6% responded to some degree. Hepatic artery ligation intraportal infusion (group IV) offered significantly longer survival than infusion via the intra-arterial route (group III) (p = 0.023). However, when compared with the control group, the survival times in the 4 treatment groups were not significantly different. Hepatic dearterialization (group H) had the highest 7-day mortality rate, 33.3%. The results of this trial suggested that none of the modalities of treatment studied offered patients with unresectable hepatocellular carcinoma any better survival and symptomatic palliation than did symptomatic treatment alone.