Clinical trials evaluating systemic therapies for hepatocellular carcinomas (HCCs) must take into consideration prognostic factors that influence survival. Clinical features such as poor performance status, older age, and the presence of jaundice may be predictive of short survivals. Because patients with elevated alpha-fetoprotein (AFP) levels have short survivals, randomized clinical trials of therapy for HCC should stratify patients based on AFP levels. Numerous clinical trials of single-agent chemotherapy for HCC have examined alkylating agents, antitumor antibiotics, anthracyclines, hormonal therapies, and interferons. However, these trials have failed to demonstrate an agent that consistently produces response rates greater than 20% or that improves survival. Similarly, effective combination chemotherapy regimens for HCC have not been identified. Clinical trials have also failed to identify effective systemic therapies for cholangiocarcinomas, although these trials have had limited numbers of patients. These dismal results mandate an examination of phase I1 agents in untreated good-performance patients with HCC. Novel approaches to drug delivery should also be pursued. 0 1993 WiIey-Liss, Inc. ~~~~