Ten patients with hepatocellular carcinoma, three of whom had pulmonary metastasis, were treated with adoptive immunotherapy using autologous lymphokine-activated killer cells plus recombinant interleukin 2. Patients received 15 micrograms per day of recombinant interleukin 2 consecutively (for 14 to 64 days), from Day 7 prior to the first leukapheresis, and received 10(9) to 10(10) lymphokine-activated killer cells once or twice per week intravenously; the lymphokine-activated killer cells had been generated from mononuclear cells obtained through leukapheresis. Preadministration of recombinant interleukin 2 prior to the first leukapheresis resulted in a remarkable increase of lymphokine-activated killer activity in seven of nine cases in whom lymphokine-activated killer activity had been poorly inducible even at high concentrations of recombinant interleukin 2. At the end of the treatment, liver tumor regression (34 and 63%, respectively, of two-dimensional size) was observed in two of two patients with a solitary tumor; no increase of liver tumor size was observed in seven patients with massive or multiple tumors, and no changes in the size or number of pulmonary metastatic tumors in any patients were observed. More than a 35% decrease in serum alpha-fetoprotein level was noted in four of nine alpha-fetoprotein-positive patients. However, Child's grades, performance status and lymphokine-activated killer activity on entry into the study could not be used as parameters to predict therapy responsiveness. Neither serious side effects nor significant changes of serum bilirubin, ALT and creatinine were noted. Thus, this treatment seems to be well tolerated even in advanced hepatocellular carcinoma with poor liver function reserve, and tumor regression could be expected in small-burden hepatocellular carcinoma. The assessment of the therapeutic effects and application in hepatocellular carcinoma awaits the development of this trial.