Immunogenic features of some malignancies have aroused interest in immunotherapy of cancer. Immunotherapy seems most effective in patients with a small tumour burden, and the focus of immunotherapy trials has, thus, lately been on adjuvant treatment. To enable further development of immunotherapy we need to know more about the mechanisms involved in host defence, especially when the system is influenced by extrinsic factors, that is, immunomodulative agents. T lymphocytes pray an important role in the host defence against tumour cells trying to escape from immune surveillance. The mechanisms that regulate the host defence systems are complex, and the influence of extrinsic factors such as immunotherapeutic agents is poorly understood. Most data on lymphocyte subsets in malignant disease originate from melanoma or renal cell carcinoma (RCC) studies, although there are scattered data on lymphocyte subsets also in other malignancies. There are several studies implying that the relative amount of CD4+, CD8+, and natural kilter (NK) cells may be important and that, by reducing the tumour burden or by using different therapeutic agents, we can stimulate the host defence. However, only some of these studies imply that these changes can have an impact on clinical outcome and prognosis. The findings of the studies reviewed in this paper are mostly encouraging, but whether the lymphocyte subsets have any value as prognostic markers in patients with malignancies receiving immunotherapy is still unclear. Large randomized immunotherapy trials including an observation arm give an ideal opportunity to recognize those immunological changes that are due to therapy, related to the natural host defence, or whether they have any prognostic value.