The dynamic interaction between the host immune system and growing cancer has been of central interest to the field of tumor immunology over the past years. Recognition of tumor-associated antigens (TAA) by self-HLA (human leukocyte antigen) class I-restricted CD8+ T cells is a main feature in the detection and destruction of malignant cells. The discovery and molecular characterization of TAA has changed the field of cancer treatment and introduced a new era of cancer immunotherapy aimed at increasing tumor immunogenicity and T-cell-mediated anti-tumor immunity. Unfortunately, while these new protocols of cancer immunotherapy are mediating induction of tumor-specific T lymphocytes in patients with certain malignancies, they have not yet delivered substantial clinical benefits, such as induction of tumor regression or increased disease-free survival. It has become apparent that lack of tumor rejection is the result of immune selection and escape by tumor cells that develop low immunogenic phenotypes. Substantial experimental data support the existence of a variety of different mechanisms involved in the tumor escape phase, including loss or downregulation of HLA class I antigens. These alterations could be caused by regulatory ('soft') or by structural/irreversible ('hard') defects. On the basis of the evidence obtained from experimental mouse cancer models and metastatic human tumors, the structural defects underlying HLA class I loss may have profound implications on T-cell-mediated tumor rejection and ultimately on the outcome of cancer immunotherapy. Strategies to overcome this obstacle, including gene therapy to recover normal expression of HLA class I genes, require consideration. In this review, we outline the importance of monitoring and correction of HLA class I alterations during cancer development and immunotherapy.