The second part of this review examines the use of recombinant interferon-alpha (rIFNalpha) in the following solid tumours: superficial bladder cancer, Kaposi's sarcoma, head and neck cancer, gastrointestinal cancers, lung cancer, mesothelioma and ovarian, breast and cervical malignancies. In superficial bladder cancer, intravesical rIFNalpha has a promising role as second-line therapy in patients resistant or intolerant to intravesical bacille Calmette-Guérin (BCG). In HIV-associated Kaposi's sarcoma, rIFNalpha is active as monotherapy and in combination with antiretroviral agents, especially in patients with CD4 counts >200/mm(3), no prior opportunistic infections and nonvisceral disease. rIFNalpha has shown encouraging results when used in combination with retinoids in the chemoprevention of head and neck squamous cell cancers. It is effective in the chemoprevention of hepatocellular cancer in hepatitis C-seropositive patients. In neuroendocrine tumours, including carcinoid tumour, low-dosage (=3 MU) or intermediate-dosage (5 to 10 MU) rIFNalpha is indicated as second-line treatment, either with octreotide or alone in patients resistant to somatostatin analogues. Intracavitary IFNalpha may be useful in malignant pleural effusions from mesothelioma. Similarly, intraperitoneal IFNalpha may have a role in the treatment of minimal residual disease in ovarian cancer. In breast cancer, the only possible role for IFNalpha appears to be intralesional administration for resistant disease. IFNalpha may have a role as a radiosensitising agent for the treatment of cervical cancer; however, this requires confirmation in randomised trials. On the basis of current evidence, the routine use of rIFNalpha is not recommended in the therapy of head and neck squamous cell cancers, upper gastrointestinal tract, colorectal and lung cancers, or mesothelioma. Pegylated IFNalpha (peginterferon-alpha) is an exciting development that offers theoretical advantages of increased efficacy, reduced toxicity and improved compliance. Further data from randomised studies in solid tumours are needed where rIFNalpha has activity, such as neuroendocrine tumours, minimal residual disease in ovarian cancer, and cervical cancer. A better understanding of the biological mechanisms that determine response to rIFNalpha is needed. Studies of IFNalpha-stimulated gene expression, which are now feasible, should help to identify molecular predictors of response and allow us to target therapy more selectively to patients with solid tumours responsive to IFNalpha.