<h2>Abstract</h2> <b>Background:</b> Because the prognosis of patients with hepatocellular carcinoma is not fully understood, particularly regarding therapy, we have evaluated it in a series of patients with a homogeneous diagnostic and therapeutic work-up. <b>Methods:</b> From 1985 to 1996, 42 variables were recorded prospectively in 178 consecutive patients who had a diagnosis of hepatocellular carcinoma. Treatment consisted of liver transplantation (n = 22), partial hepatectomy (n = 11), arterial chemoembolization (n = 52), systemic or regional chemotherapy (n = 51), and other therapies (n = 5); 37 patients received no specific therapy. Statistical analysis was performed according to a Cox model. <b>Results:</b> There were no differences between the survival of patients receiving chemotherapy, other therapies, or no treatment (control group, n=93). Survival rates at 1, 3, and 5 years were 81%, 74%, and 74% for liver transplantation, 72%, 58%, and 58% for hepatectomy, 55%, 26%, and 13% for chemoembolization, and 13%, 3%, and 0% for the control group. Cirrhosis, systemic syndrome, bilobar involvement, Child's stage C disease, and treatment were independent predictors of survival. <b>Conclusions:</b> This series shows that certain easily accessible parameters may help establish individual prognosis and stratify patients in clinical trials and indicates that chemoembolization, partial resection, and liver transplantation can prolong life expectancy of patients with hepatocellular carcinoma. (Surgery 1998;124:575-83.)