The identification of tumor-associated antigens recognized by cellular or humoral effectors of the immune system has opened new perspectives for cancer therapy. Different groups of cancer-associated antigens have been described as targets for cytotoxic T lymphocytes (CTLs) in vitro and in vivo: 1) cancer-testis (CT) antigens, which are expressed in different tumors and normal testis; 2) melanocyte differentiation antigens; 3) point mutations of normal genes; 4) antigens that are overexpressed in malignant tissues; and 5) viral antigens. Clinical studies with peptides derived from these antigens have been initiated to induce specific CTL responses in vivo. Immunological and clinical parameters for the assessment of peptide-specific reactions have been defined, i.e., delayed-type hypersensitivity (DTH), CTL, autoimmmune, and tumor regression responses. Preliminary results demonstrate that tumor-associated peptides alone elicit specific DTH and CTL responses leading to tumor regression after intradermal injection. Granulocyte-macrophage colony-stimulating factor (GM-CSF) was proven effective in enhancing peptide-specific immune reactions by amplification of dermal peptide-presenting dendritic cells. Long-lasting complete tumor regressions have been observed after induction of peptide-specific CTLs. However, in single cases with disease progression after an initial tumor response, either a loss of the respective tumor antigen targeted by CTLs or of the presenting major histocompatibility complex (MHC) class I allele was detected as a mechanism of immune escape under immunization. Based on these observations, cytokines to enhance antigen and MHC class I expression in vivo are being evaluated to prevent immunoselection. Recently, a strategy utilizing spontaneous antibody responses to tumor-associated antigens (SEREX) has led to the identification of a new CT antigen, NY-ESO-1, which is regarded as one of the most immunogenic antigens known today inducing spontaneous immune responses in 50% of patients with NY-ESO-1-expressing cancers. Clinical studies involving antigenic constructs that induce both antibody and CTL responses will show whether these are more effective for immunotherapy of cancer.